Fit the questionnaire seed into the measured comp pipeline, fix CI

The live-reference test still asserted SKILL.src.md's pre-rewrite setup phrasing, which failed every pipeline on main and here and blocked the generated-output sync; the assertions now match the current text. The sync list gains .veto, added in #675 but never listed. Two seed-world seams in the comp-led phases close: DESIGN.md's named faces ride into font-match as candidates instead of losing silently to a metric lookalike, and a sampled spec color within render noise of a recorded hex builds at the recorded value. The cue also stops claiming type character it cannot carry; comp prompts name the recorded faces. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Restore generated provider output to main's state
2026-09-11 21:57:14 +03:00 · 2026-09-01 15:55:54 +05:00 · 2026-09-01 10:21:53 +05:00 · 2026-09-01 10:18:14 +05:00 · 2026-09-01 10:13:02 +05:00 · 2026-09-01 10:09:45 +05:00
299 changed files with 49470 additions and 570 deletions
@@ -14,7 +14,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .agents/skills/impeccable/scripts/...` command in this skill and its references, and `.agents/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -97,7 +97,7 @@ Codex failure accounting: final Run Notes must include target slug, ignore list,

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Codex final-answer note: `$impeccable critique` produces a report artifact, so the final chat response should intentionally exceed the usual concise close-out style. Do not title the final response "Critique Summary" unless the user explicitly asked for a summary.

@@ -219,7 +219,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .agents/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .agents/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .agents/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .agents/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .agents/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .agents/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .agents/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `$impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -2,7 +2,7 @@
  "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
  "name": "impeccable",
  "metadata": {
-    "description": "Design fluency for AI harnesses. 1 skill, 23 commands, and curated anti-patterns for impeccable frontend design."
+    "description": "Design fluency for AI harnesses. 1 skill, 24 commands, and curated anti-patterns for impeccable frontend design."
  },
  "owner": {
    "name": "Paul Bakaus",
@@ -11,7 +11,7 @@
  "plugins": [
    {
      "name": "impeccable",
-      "description": "Design fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.",
+      "description": "Design fluency for frontend development. 1 skill with 24 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.",
      "version": "4.1.2",
      "author": {
        "name": "Paul Bakaus",
@@ -1,6 +1,6 @@
 {
  "name": "impeccable",
-  "description": "Design fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.",
+  "description": "Design fluency for frontend development. 1 skill with 24 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.",
  "version": "4.1.2",
  "author": {
    "name": "Paul Bakaus",
@@ -20,7 +20,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .claude/skills/impeccable/scripts/...` command in this skill and its references, and `.claude/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .claude/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .claude/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .claude/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .claude/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .claude/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .claude/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .claude/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -15,7 +15,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .cursor/skills/impeccable/scripts/...` command in this skill and its references, and `.cursor/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .cursor/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .cursor/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .cursor/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .cursor/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .cursor/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .cursor/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .cursor/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -14,7 +14,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .gemini/skills/impeccable/scripts/...` command in this skill and its references, and `.gemini/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .gemini/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .gemini/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .gemini/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .gemini/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .gemini/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .gemini/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .gemini/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -17,7 +17,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .github/skills/impeccable/scripts/...` command in this skill and its references, and `.github/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .github/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .github/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .github/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .github/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .github/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .github/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .github/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -21,10 +21,12 @@ concurrency:

 env:
  GENERATED_PATHS: >-
+    .agent
    .agents
    .claude
    .cursor
    .gemini
+    .github/agents
    .github/skills
    .grok
    .hermes
@@ -35,6 +37,7 @@ env:
    .rovodev
    .trae
    .trae-cn
+    .veto
    .vibe
    plugin

@@ -61,6 +61,14 @@ Thumbs.db
 .impeccable/live/deferred-svelte-component-accepts.json
 .impeccable/history/
 .impeccable/config.local.json
+# API keys for optional image-generation cues (document seed mode).
+.impeccable/.env
+# Generated visual-cue images + cues.json (document seed Step 4).
+.impeccable/visual-cues/
+# The design-context store itself is the user's own record and is trackable;
+# only its session state and its generated exports stay local.
+.impeccable/design-context/runtime/
+.impeccable/design-context/exports/
 **/.impeccable/hook.pending.json
 .impeccable/provider-smoke/
 src/__impeccable_provider_smoke_*.html
@@ -129,3 +137,9 @@ tmp/
 # PNGs, the old card backup). The canonical generator is `bun run og-image`
 # (scripts/generate-og-image.js); this dir is throwaway and safe to delete.
 .og-build/
+
+
+# Intermediate and generated output for the picker build. Release builds sync
+# the generated picker into the tracked provider harness directories.
+build-picker/
+skill/scripts/picker/
@@ -20,7 +20,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .grok/skills/impeccable/scripts/...` command in this skill and its references, and `.grok/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .grok/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .grok/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .grok/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .grok/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .grok/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .grok/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .grok/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -15,7 +15,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .hermes/skills/impeccable/scripts/...` command in this skill and its references, and `.hermes/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .hermes/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .hermes/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .hermes/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .hermes/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .hermes/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .hermes/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .hermes/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -4,7 +4,8 @@
    "ignoreFiles": [
      "tests/fixtures/**",
      "tests/framework-fixtures/**",
-      "tests/detect-antipatterns.test.js"
+      "tests/detect-antipatterns.test.js",
+      "picker/styles/design-context.css"
    ],
    "ignoreValues": [
      {
@@ -15,6 +16,41 @@
        ],
        "createdAt": "2026-07-17T00:00:00.000Z",
        "reason": "Live overlay chrome is injected over arbitrary host pages and builds a self-contained UI with its own small type scale; DESIGN.md's ramp describes the impeccable website, not this widget"
+      },
+      {
+        "rule": "broken-image",
+        "value": "*",
+        "files": [
+          "picker/pages/index.astro"
+        ],
+        "createdAt": "2026-07-23T14:51:18.241Z"
+      },
+      {
+        "rule": "design-system-font-size",
+        "value": "*",
+        "files": [
+          "picker/styles/picker.css"
+        ],
+        "createdAt": "2026-07-23T21:00:00.000Z",
+        "reason": "The picker is a self-contained questionnaire widget with its own kit-token type scale, and screen 04's cards render arbitrary Google Fonts at specimen sizes; DESIGN.md's ramp describes the impeccable website, not this tool (same footing as the live-browser.js entry)"
+      },
+      {
+        "rule": "design-system-font",
+        "value": "*",
+        "files": [
+          "picker/scripts/design-context.js"
+        ],
+        "createdAt": "2026-07-30T13:11:54.730Z",
+        "reason": "Runtime font loader: builds a Google Fonts URL from the family the agent/user chose in the questionnaire; no literal face is being introduced"
+      },
+      {
+        "rule": "bounce-easing",
+        "value": "cubic-bezier(.34, 1.56, .64, 1)",
+        "files": [
+          "picker/scripts/palette-picker.js"
+        ],
+        "createdAt": "2026-08-20T02:53:58.453Z",
+        "reason": "plan-18 step 10c: byte-exact port from tmp/motion-previews-premium.html installChoreographedPremium spring stagger"
      }
    ]
  },
@@ -15,7 +15,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .kiro/skills/impeccable/scripts/...` command in this skill and its references, and `.kiro/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .kiro/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .kiro/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .kiro/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .kiro/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .kiro/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .kiro/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .kiro/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -20,7 +20,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .opencode/skills/impeccable/scripts/...` command in this skill and its references, and `.opencode/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .opencode/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .opencode/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .opencode/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .opencode/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .opencode/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .opencode/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .opencode/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
@@ -91,7 +91,7 @@

 import crypto from 'node:crypto';
 import { dirname, join, relative, resolve } from 'node:path';
-import { readFileSync } from 'node:fs';
+import { readFileSync, realpathSync } from 'node:fs';
 import { fileURLToPath } from 'node:url';
 import {
  approvedPoolRevision,
@@ -703,7 +703,28 @@ export function nextStepAfterChoice({ key, scope, cwd = process.cwd(), env = pro
  return `NEXT (comp-led, ${why}): the world is chosen; the composition is not. Run: node ${scripts}/build-phase.mjs start${seed} and follow its NEXT lines: it opens the comps phase (three comps under .impeccable/mocks/, one approved by the user through the decision page or structured question, sidecar "approved": true), then spec, plates, hero, sections, motion, responsive, review. Do not write page code before those gates close. Reference: reference/visualize.md for the comp round.\n`;
 }

-if (process.argv[1] && resolve(process.argv[1]) === fileURLToPath(import.meta.url)) {
+export function sameMainModulePath(left, right, platform = process.platform) {
+  if (platform !== 'win32') return left === right;
+  const normalizeDriveLetter = (value) => value.replace(/^([a-z]):/i, (_, drive) => `${drive.toUpperCase()}:`);
+  return normalizeDriveLetter(left) === normalizeDriveLetter(right);
+}
+
+function isMainModule() {
+  if (!process.argv[1]) return false;
+  try {
+    // Node resolves import.meta.url through symlinks but leaves argv[1] as the
+    // invoked path. Compare real paths so a linked skill still runs its CLI,
+    // normalizing the drive-letter casing that Windows junctions can change.
+    return sameMainModulePath(
+      realpathSync(process.argv[1]),
+      realpathSync(fileURLToPath(import.meta.url))
+    );
+  } catch {
+    return false;
+  }
+}
+
+if (isMainModule()) {
  const args = process.argv.slice(2);
  const fromIdx = args.indexOf('--from');
  const scopeIdx = args.indexOf('--scope');
@@ -16,8 +16,9 @@
 * CLI entry points (called from skill instructions):
 *   node critique-storage.mjs slug <resolved-target>
 *   node critique-storage.mjs write <slug> <snapshot-body-file>
- *   node critique-storage.mjs latest <slug>
+ *   node critique-storage.mjs latest <slug> [--json]
 *   node critique-storage.mjs trend <slug> [limit]
+ *   node critique-storage.mjs close <resolved-target> <snapshot-file>
 *
 * Note: there is intentionally no `ignore` subcommand. ignore.md is a plain
 * markdown file; the model reads it directly with its file-read tool. This
@@ -27,6 +28,7 @@

 import fs from 'node:fs';
 import path from 'node:path';
+import { createHash } from 'node:crypto';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { getCritiqueDir } from './lib/impeccable-paths.mjs';
 import { slugFromTarget } from './lib/target-slug.mjs';
@@ -50,6 +52,45 @@ export function nowFilenameStamp(date = new Date()) {
  return iso.replace(/[:.]/g, '-').replace(/-\d+Z$/, 'Z');
 }

+/**
+ * Return an exact content fingerprint for a local file target. URLs and
+ * non-files return null because their content is not available here.
+ *
+ * The fingerprint deliberately describes bytes, not Git state or mtimes:
+ * critique often assesses an uncommitted file, and a later polish run should
+ * inherit that backlog when the bytes are unchanged regardless of staging.
+ */
+function resolveLocalTargetPath(target, { cwd = process.cwd() } = {}) {
+  if (!target || /^https?:\/\//i.test(target)) return null;
+  return path.isAbsolute(target) ? path.resolve(target) : path.resolve(cwd, target);
+}
+
+function resolveTargetIdentity(target, { cwd = process.cwd() } = {}) {
+  if (!target || typeof target !== 'string') return null;
+  if (/^https?:\/\//i.test(target)) {
+    try {
+      const url = new URL(target);
+      const pathname = url.pathname.replace(/\/+$/, '') || '/';
+      return `url:${url.origin}${pathname}`;
+    } catch {
+      return null;
+    }
+  }
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  return filePath ? `file:${filePath}` : null;
+}
+
+export function fingerprintTarget(target, { cwd = process.cwd() } = {}) {
+  const filePath = resolveLocalTargetPath(target, { cwd });
+  if (!filePath) return null;
+  try {
+    if (!fs.statSync(filePath).isFile()) return null;
+    return `sha256:${createHash('sha256').update(fs.readFileSync(filePath)).digest('hex')}`;
+  } catch {
+    return null;
+  }
+}
+
 /**
 * Write a snapshot for `slug`. `meta` carries the small structured frontmatter
 * keys read back by readTrend(). `body` is the human-readable critique
@@ -62,14 +103,27 @@ export function writeSnapshot({ slug, meta, body, cwd = process.cwd(), now = new
  const dir = getCritiqueDir(cwd);
  fs.mkdirSync(dir, { recursive: true });
  const timestamp = nowFilenameStamp(now);
-  const filePath = path.join(dir, `${timestamp}__${slug}.md`);
  // Spread `meta` first so internally computed `timestamp` and `slug`
  // always win. Otherwise a caller-supplied meta blob (parsed from the
  // IMPECCABLE_CRITIQUE_META env var) could clobber them, leaving the
  // filename in disagreement with its frontmatter and corrupting trends.
  const front = serializeFrontmatter({ ...meta, timestamp, slug });
-  fs.writeFileSync(filePath, `${front}\n${body.trim()}\n`, 'utf-8');
-  return filePath;
+  const contents = `${front}\n${body.trim()}\n`;
+
+  // A second critique can finish in the same UTC second. Use exclusive
+  // creation and a fixed-width suffix so concurrent writers cannot replace
+  // history and lexical ordering still keeps collision entries newest.
+  for (let collision = 0; collision <= 9999; collision += 1) {
+    const suffix = collision === 0 ? '' : `~${String(collision).padStart(4, '0')}`;
+    const filePath = path.join(dir, `${timestamp}${suffix}__${slug}.md`);
+    try {
+      fs.writeFileSync(filePath, contents, { encoding: 'utf-8', flag: 'wx' });
+      return filePath;
+    } catch (error) {
+      if (error?.code !== 'EEXIST') throw error;
+    }
+  }
+  throw new Error(`Too many critique snapshots for ${slug} at ${timestamp}`);
 }

 function serializeFrontmatter(obj) {
@@ -98,6 +152,8 @@ function parseFrontmatter(text) {
      try { value = JSON.parse(value); } catch { /* leave as-is */ }
    } else if (/^-?\d+$/.test(value)) {
      value = Number(value);
+    } else if (value === 'true' || value === 'false') {
+      value = value === 'true';
    }
    out[key] = value;
  }
@@ -107,7 +163,7 @@ function parseFrontmatter(text) {
 /**
 * Return snapshot files matching `suffix`, sorted oldest → newest.
 */
-const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z__.+\.md$/;
+const SNAPSHOT_FILENAME = /^\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}Z(?:~\d{4})?__.+\.md$/;

 function listSnapshots(suffix, cwd) {
  const dir = getCritiqueDir(cwd);
@@ -118,24 +174,105 @@ function listSnapshots(suffix, cwd) {
    .map((f) => path.join(dir, f));
 }

-function readLatestSnapshotMatching(suffix, cwd) {
-  const filePath = listSnapshots(suffix, cwd).at(-1);
+function readSnapshot(filePath) {
  if (!filePath) return null;
  const body = fs.readFileSync(filePath, 'utf-8');
  return { path: filePath, body, meta: parseFrontmatter(body) };
 }

+function snapshotTargetIdentity(snapshot) {
+  const targetPath = snapshot?.meta.target_path;
+  return snapshot?.meta.target_identity
+    || (targetPath ? `file:${targetPath}` : null);
+}
+
+function readNewestSnapshot(slug, { cwd = process.cwd() } = {}) {
+  return readSnapshot(listSnapshots(`__${slug}.md`, cwd).at(-1));
+}
+
+function readNewestSnapshotForIdentity(
+  slug,
+  targetIdentity,
+  { cwd = process.cwd() } = {},
+) {
+  const matches = listSnapshots(`__${slug}.md`, cwd)
+    .map(readSnapshot)
+    .filter((snapshot) => snapshotTargetIdentity(snapshot) === targetIdentity);
+  return matches.at(-1) || null;
+}
+
 /**
 * Return the most recent snapshot for `slug`, or null. Polish reads this
 * to find its fix backlog when the slug matches.
 */
 export function readLatestSnapshot(slug, { cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching(`__${slug}.md`, cwd);
+  const latest = readNewestSnapshot(slug, { cwd });
+  return latest?.meta.closed === true ? null : latest;
+}
+
+/**
+ * Mark one exact snapshot closed without deleting the score history consumed
+ * by `trend`. Exact identity matters: a newer critique may land after polish
+ * reads its backlog, and that newer snapshot must remain live. `snapshotFile`
+ * may be the absolute path returned by readLatestSnapshot() or the basename
+ * emitted by `latest --json`. Returns the path marked closed, or null.
+ */
+export function closeSnapshot(snapshotFile, { cwd = process.cwd() } = {}) {
+  if (!snapshotFile || typeof snapshotFile !== 'string') return null;
+  const dir = path.resolve(getCritiqueDir(cwd));
+  const snapshotPath = path.isAbsolute(snapshotFile)
+    ? path.resolve(snapshotFile)
+    : path.resolve(dir, snapshotFile);
+  const filename = path.basename(snapshotPath);
+  if (
+    path.dirname(snapshotPath) !== dir
+    || !SNAPSHOT_FILENAME.test(filename)
+  ) return null;
+
+  let snapshot;
+  try {
+    if (!fs.lstatSync(snapshotPath).isFile()) return null;
+    snapshot = readSnapshot(snapshotPath);
+  } catch {
+    return null;
+  }
+  if (!snapshot || snapshot.meta.closed === true) return null;
+  const closedBody = snapshot.body.replace(
+    /^(---\r?\n[\s\S]*?)(\r?\n---)/,
+    '$1\nclosed: true$2',
+  );
+  if (closedBody === snapshot.body) {
+    throw new Error(`Cannot close snapshot without frontmatter: ${snapshot.path}`);
+  }
+  fs.writeFileSync(snapshot.path, closedBody, 'utf-8');
+  return snapshot.path;
 }

 /** Return the most recent snapshot across all targets, or null. */
 export function readLatestSnapshotAcrossTargets({ cwd = process.cwd() } = {}) {
-  return readLatestSnapshotMatching('.md', cwd);
+  const snapshots = listSnapshots('.md', cwd).map(readSnapshot);
+  const identifiedSlugs = new Set(
+    snapshots
+      .filter((snapshot) => snapshotTargetIdentity(snapshot))
+      .map((snapshot) => snapshot.meta.slug),
+  );
+  const latestByTarget = new Map();
+  for (const snapshot of snapshots) {
+    if (!snapshot?.meta.slug) continue;
+    // Slugs are lossy: distinct targets such as foo/bar and foo-bar can share
+    // one. Keep each known identity's latest open/closed state independent so
+    // closing one target cannot hide another target's live backlog. Once a
+    // slug has any identity-aware snapshot, its older legacy records are no
+    // longer independently routable and must not resurface as zombie work.
+    const targetIdentity = snapshotTargetIdentity(snapshot);
+    if (!targetIdentity && identifiedSlugs.has(snapshot.meta.slug)) continue;
+    const streamKey = targetIdentity || `slug:${snapshot.meta.slug}`;
+    latestByTarget.set(streamKey, snapshot);
+  }
+  return [...latestByTarget.values()]
+    .filter((snapshot) => snapshot.meta.closed !== true)
+    .sort((a, b) => a.path.localeCompare(b.path))
+    .at(-1) || null;
 }

 /**
@@ -153,9 +290,13 @@ export function readTrend(slug, { limit = 5, cwd = process.cwd() } = {}) {
 // Accept either a ready slug or a concrete target (path/URL) everywhere, so
 // callers never have to run the slug step separately. Anything containing a
 // path or URL marker is resolved through slugFromTarget.
+function isReadySlug(value) {
+  return /^[a-z0-9-]+$/.test(value || '') && !value.includes('/');
+}
+
 function coerceSlug(value) {
  if (!value) return null;
-  if (/^[a-z0-9-]+$/.test(value) && !value.includes('/')) return value;
+  if (isReadySlug(value)) return value;
  return slugFromTarget(value);
 }

@@ -181,14 +322,124 @@ function main(argv) {
      if (metaArg) {
        try { meta = JSON.parse(metaArg); } catch { /* ignore */ }
      }
+      // The helper, not caller-provided metadata, owns the target fingerprint.
+      // This makes the snapshot describe the exact file bytes critique saw.
+      delete meta.target_fingerprint;
+      delete meta.target_path;
+      delete meta.target_identity;
+      const targetIdentity = resolveTargetIdentity(slugArg);
+      if (targetIdentity) meta.target_identity = targetIdentity;
+      const targetFingerprint = fingerprintTarget(slugArg);
+      if (targetFingerprint) {
+        meta.target_fingerprint = targetFingerprint;
+        meta.target_path = resolveLocalTargetPath(slugArg);
+      }
      const out = writeSnapshot({ slug, meta, body: raw });
      process.stdout.write(`${out}\n`);
      return;
    }
    case 'latest': {
-      const latest = readLatestSnapshot(coerceSlug(args[0]));
-      if (!latest) { process.exit(2); }
-      process.stdout.write(latest.body);
+      const target = args[0];
+      const format = args[1];
+      const slug = coerceSlug(target);
+      if (!slug || (format && format !== '--json')) {
+        process.stderr.write('usage: latest <slug-or-target> [--json]\n');
+        process.exit(1);
+      }
+      const targetFingerprint = fingerprintTarget(target);
+      const targetPath = resolveLocalTargetPath(target);
+      const targetIdentity = resolveTargetIdentity(target);
+      const readySlug = isReadySlug(target);
+      const newestForSlug = readNewestSnapshot(slug);
+      if (!newestForSlug) { process.exit(2); }
+
+      // Concrete targets select the newest snapshot for their exact identity,
+      // not merely the newest filename for a lossy slug. This keeps distinct
+      // targets such as foo/bar and foo-bar from hiding each other's backlog.
+      const exactSnapshot = readNewestSnapshotForIdentity(slug, targetIdentity);
+      let latest = exactSnapshot;
+      if (!latest && !readySlug) {
+        // Legacy snapshots have no identity. Preserve their old explicit
+        // path/URL behavior only when no known target identity was selected.
+        latest = readNewestSnapshotForIdentity(slug, null);
+      }
+      if (!latest) latest = newestForSlug;
+      if (latest.meta.closed === true) { process.exit(2); }
+
+      const recordedTargetPath = latest.meta.target_path;
+      const recordedTargetIdentity = snapshotTargetIdentity(latest);
+      const matchingIdentity = recordedTargetIdentity === targetIdentity;
+
+      // Bare slugs remain a supported lookup mode, including for URL
+      // snapshots. But when a same-named local file exists, the request is
+      // ambiguous unless that exact file owns the snapshot identity.
+      if (readySlug && !recordedTargetIdentity) {
+        process.stderr.write(
+          'ambiguous legacy snapshot target; use an explicit ./path or full URL\n',
+        );
+        process.exit(2);
+      }
+      if (readySlug && targetPath && fs.existsSync(targetPath) && !matchingIdentity) {
+        process.stderr.write(
+          'ambiguous snapshot slug; use an explicit ./path or remove the local name collision\n',
+        );
+        process.exit(2);
+      }
+
+      const concreteTarget = !readySlug || matchingIdentity;
+      if (concreteTarget && recordedTargetIdentity && !matchingIdentity) {
+        process.exit(2);
+      }
+      const concreteLocalTarget = concreteTarget && targetPath;
+      if (concreteLocalTarget && latest.meta.target_fingerprint !== targetFingerprint) {
+        closeSnapshot(latest.path);
+        process.exit(2);
+      }
+      if (format === '--json') {
+        process.stdout.write(JSON.stringify({
+          snapshot_file: path.basename(latest.path),
+          body: latest.body,
+        }, null, 2) + '\n');
+      } else {
+        process.stdout.write(latest.body);
+      }
+      return;
+    }
+    case 'close': {
+      const [slugArg, snapshotFile, ...extra] = args;
+      const slug = coerceSlug(slugArg);
+      if (!slug || !snapshotFile || extra.length > 0) {
+        process.stderr.write('usage: close <resolved-target> <snapshot-file>\n');
+        process.exit(1);
+      }
+      if (
+        path.basename(snapshotFile) !== snapshotFile
+        || !SNAPSHOT_FILENAME.test(snapshotFile)
+        || !snapshotFile.endsWith(`__${slug}.md`)
+      ) process.exit(2);
+
+      // A slug and filename are not enough to prove ownership because two
+      // distinct targets can normalize to the same slug. Modern snapshots
+      // carry a canonical identity, so require the supplied resolved target
+      // to match it before allowing the exact snapshot to be closed. Legacy
+      // snapshots without identity retain their historical close behavior.
+      const snapshotPath = path.join(getCritiqueDir(process.cwd()), snapshotFile);
+      let snapshot;
+      try {
+        if (!fs.lstatSync(snapshotPath).isFile()) process.exit(2);
+        snapshot = readSnapshot(snapshotPath);
+      } catch {
+        process.exit(2);
+      }
+      const recordedTargetIdentity = snapshotTargetIdentity(snapshot);
+      if (
+        recordedTargetIdentity
+        && recordedTargetIdentity !== resolveTargetIdentity(slugArg)
+      ) process.exit(2);
+
+      const closed = closeSnapshot(snapshotFile);
+      if (!closed) { process.exit(2); }
+      process.stdout.write(`${closed}\n`);
      return;
    }
    case 'trend': {
@@ -197,7 +448,7 @@ function main(argv) {
      return;
    }
    default:
-      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend> [args]\n');
+      process.stderr.write('usage: critique-storage.mjs <slug|write|latest|trend|close> [args]\n');
      process.exit(1);
  }
 }
@@ -995,7 +995,9 @@ function extractRadiusTokens(value) {
  return String(value || '')
    .replace(/\s*\/\s*/g, ' ')
    .split(/\s+/)
-    .map(token => token.trim())
+    // var() fallbacks leave the closing parenthesis on the final token. Strip
+    // it before length resolution so `8px)` is not treated as unitless 8rem.
+    .map(token => token.trim().replace(/\)+$/, ''))
    .filter(Boolean);
 }

@@ -431,12 +431,15 @@ fs.writeFileSync(out, Buffer.from(b64, 'base64'));
 // The prompt travels with the asset: embedded in the file itself (EXIF-class
 // metadata via embed-prompt.mjs) so intent survives copies across harnesses,
 // plus a sidecar for anything that indexes rather than opens the image.
+let embedded = false;
 try {
  const { spawnSync } = await import('node:child_process');
-  spawnSync(process.execPath, [new URL('./embed-prompt.mjs', import.meta.url).pathname, out, '--prompt', prompt], { stdio: 'ignore' });
+  const result = spawnSync(process.execPath, [fileURLToPath(new URL('./embed-prompt.mjs', import.meta.url)), out, '--prompt', prompt], { stdio: 'ignore' });
+  embedded = !result.error && result.status === 0;
+  if (!embedded) console.warn('generate-image: failed to embed prompt in the image');
  fs.writeFileSync(`${out}.json`, JSON.stringify({ prompt, createdAt: new Date().toISOString(), tool: 'generate-image.mjs', model: 'gpt-image-2', ...(refs.length ? { refs } : {}) }, null, 2));
 } catch { /* embedding is best-effort */ }
-console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); prompt embedded + sidecar at ${out}.json`);
+console.log(`IMAGE: ${out} (${size}, ${quality}, gpt-image-2, billed to your OpenAI key); ${embedded ? 'prompt embedded + sidecar' : 'sidecar'} at ${out}.json`);
 if (plateCtx && plateCtx.chroma) {
  const frac = await keyChroma(out, plateCtx.chroma);
  console.log(`PLATE-CHROMA keyed ${(frac * 100).toFixed(0)}% of pixels to alpha (${plateCtx.chroma}); place with a plain <img> over the page's own ground, no background on the plate. If the keyed fraction is under 20% the generator ignored the key: regenerate with --no-chroma and use mix-blend-mode: multiply instead.`);
@@ -768,7 +768,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
        }),
      });
      res.writeHead(200, {
-        'Content-Type': 'application/javascript',
+        'Content-Type': 'application/javascript; charset=utf-8',
        'Cache-Control': 'no-store, no-cache, must-revalidate, max-age=0',
        'Pragma': 'no-cache',
      });
@@ -777,7 +777,7 @@ function createRequestHandler({ detectScript, liveScriptParts }) {
    }
    if (p === '/detect.js' || p === '/') {
      if (!detectScript) { res.writeHead(404); res.end('Not available'); return; }
-      res.writeHead(200, { 'Content-Type': 'application/javascript' });
+      res.writeHead(200, { 'Content-Type': 'application/javascript; charset=utf-8' });
      res.end(detectScript);
      return;
    }
@@ -18,7 +18,7 @@ Core principles:
 ## Setup

 1. Run `node <skill-base-dir>/scripts/context.mjs` once per session, where `<skill-base-dir>` is the loaded base directory the runtime reports for this skill; keep cwd at the user's project. That base directory resolves every `node .pi/skills/impeccable/scripts/...` command in this skill and its references, and `.pi/skills/impeccable/scripts` is the fallback only when the runtime reports no base directory. Pass a named source file or route as `--target <path>`. It loads PRODUCT.md, DESIGN.md, the matching surface brief, and native-platform guidance when applicable; follow its directives and do not rerun it.
-2. Before acting, load the one playbook that owns the request: the Commands table's reference for an explicit or clearly implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Then inspect the target and at least one representative source of incumbent visual truth (tokens, theme, CSS, component, or asset) before editing.
+2. Load the request's playbook: its Commands-table reference for an explicit/implied sub-command, or [reference/new-work.md](reference/new-work.md) for a new surface or replacement visual world. Inspect target and incumbent visual truth before editing. When the app cannot run, start with committed visual-regression goldens or screenshot fixtures; verify target and freshness against current tokens, CSS, components, or assets, resolve conflicts, and compare theme/variant captures.
 3. After analysis and direction are resolved, load [reference/craft-floor.md](reference/craft-floor.md) immediately before editing UI. It carries the quality floor, the absolute bans, and the reflexes no detector catches. Do not load it for planning-only work.

 ## How to design
@@ -1,6 +1,6 @@
 ### Purpose

-Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive/backlog for future commands.
+Resolve one stable target, run two independent assessments, synthesize a design critique, persist a snapshot, and ask the user what to improve next. The chat response is the primary deliverable; the snapshot is an archive of that run.

 ### Hard Invariants

@@ -84,7 +84,7 @@ After Assessment B returns usable CLI findings, reuse them. Do not rerun `detect

 Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.

-The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is only an archive/backlog for later commands.
+The chat response is the primary user-facing deliverable. Present the full structured critique below in chat; do not replace it with a summary and a link. The persisted snapshot is an archive of that run.

 Structure your feedback as a design director would:

@@ -197,7 +197,7 @@ Skip this step if the Setup slug was null (vague or root-level target).
   IMPECCABLE_CRITIQUE_META='{"target":"<user phrasing>","total_score":<n>,"max_score":<n>,"na_heuristics":"<comma-separated numbers, or empty>","p0_count":<n>,"p1_count":<n>}' \
     node .pi/skills/impeccable/scripts/critique-storage.mjs write "<resolved target>" <body-file>
   ```
-   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. The helper prints the absolute path it wrote.
+   `max_score` is the applicable maximum from the heuristic table (40 when every heuristic applied), so a later run can tell a renormalized total from a full one. For a local file target, the helper also records an exact content fingerprint so polish can distinguish the assessed bytes from later edits without relying on Git state or timestamps. The helper prints the absolute path it wrote. Leave that file on disk. Polish closes it; this run does not.

 3. **Delete the temp body file** after the write attempt completes, whether the write succeeded or failed. If deletion fails, mention `temp-file cleanup failed: <reason>` briefly in the final output, but do not block the critique.

@@ -46,7 +46,7 @@ The script deals three of your structures; the dice pick which three reach the u
 4. Run `node .pi/skills/impeccable/scripts/concept-seed.mjs --scope direction --mode <mode>` and follow what it prints. No substitute, no skip: on a new or replacement world, writing artifact code before this script has run and its assignment is acknowledged is a contract violation, whatever the harness, the model, or the time pressure; the roll is what keeps every run from converging on the category default. The script assigns the direction to build and deals catalog challengers. Fuse each challenger before judging it: the challenger supplies the form and its system grammar, the product supplies every fact, clarity wins conflicts. Weigh fused challengers against the assigned direction on exactly two axes, audience identification and product clarity. Losing to strong grounded material is a valid outcome; beating a thin or tool-monoculture list is the point. Close with a verdict per challenger, decided before any borrowing: wins (beats the assigned direction on both axes; becomes the build candidate), competitive (holds one axis; stays a full alternate), or declined (loses both). A declined challenger is not spent: name the one discipline of its system the assigned direction lacks, and raise the assigned direction to match before presenting it. A donation transfers ambition and system discipline (a palette's total commitment, a grid's density courage, a form's structural honesty), never the challenger's clothes; a lifted motif is a costume note, not a raise, and one world owns the page. Write each raise into the presented direction as its own line, named for its donor; a raise nobody can read did not happen.
 5. Present one direction, fully committed and already raised by the hand it beat, raises visible as named lines: world, first viewport, visitor path, signature interaction, cross-surface reach, honest risk. Route each challenger by verdict: winning and competitive challengers are full alternates with their QUALITY BAR cards and one-line case; declined challengers render demoted, compact and quiet, each carrying its verdict and what the direction kept from it, never full-size, never silently dropped, still adoptable on request. The verdict informs the user's choice, never pre-empts it; the demoted row is the hand's proof of judgment. A hand holds at most three full-card challengers: when the roll deals more, the three strongest join and the rest wait in the re-roll pool, noted in one line; dropping a challenger from the hand itself takes a named product-truth failure, disclosed. Add one card for your own top-ranked grounded candidate when it is not the assigned direction, kicker IMPECCABLE’S PICK, same anatomy as every card, with an honest risk line naming its familiarity when true: the strongest grounded direction is often where most runs in this category land, and the user deciding that trade is the point of showing it. Familiar and effective is a legitimate destination, not a failure of nerve; the pick card and the standing exit serve it at two depths. One pick card, never two, never a ranked list: a lineup of your candidates hands selection back to a taste function and invites the safest card. The pick never takes the lead position; when the dice assign your top candidate there is no pick card, and the assigned card notes it topped your list. Add re-roll with an optional one-line steer, in three registers: plain (a fresh hand, same spread), safer (your remaining conventional grounded candidates plus the canon against named competitors), bolder (foreign forms only, at full commitment). The register is the user's steering on the familiar-to-bold axis, never yours to pre-select; when the answer carries one, re-run the seed with `--register <value>` and the next `--reroll` round, and follow what it prints. A user saying "bolder" or "safer" while a direction round is open means these registers, never the bolder or harden commands. The two channels share this structure and differ only in richness: cards and boards on the decision page, names and one-liners through the structured tool, whose option list carries the assigned direction, the pick, the winning and competitive challengers, and the standing exit last; declined challengers fold into the assigned option's description as their kept lines, so the raise survives the text channel.

-The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .pi/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.
+The standing exit: every direction round offers one quiet, permanent alternative, the category standard, played straight. It is the user's door, never yours: never recommend it, never weigh it against the roll, never let it soften the dealt directions; the counterweights bind the unchosen default, not the chosen one. When the user takes it (the canon action, a safer-steer, or plain words asking for the familiar or competitor-like path), convention becomes the commitment: ask once for two or three products this should sit alongside, make their craft level the bar, and execute the canon at full fidelity, without irony or smuggled quirk. Record a standing preference as a brand commitment in PRODUCT.md. Re-roll eliminates every direction already shown, grounded and challenger alike; after two consecutive re-rolls, ask what quality is missing. Re-roll on your own only on named factual grounds, when the assigned direction cannot carry the product's truth or task; taste is never grounds. The user may re-roll freely, and a user- or brief-pinned direction beats the roll, always. Resolve collisions field by field: preserve every user- or brief-pinned constraint. In dimensions the brief leaves open, the assignment still binds through its topology, controls, state vocabulary, and ritual; when only its materials conflict with a pinned visual direction or PRODUCT.md brand commitment, translate that material expression and name the translation in the presented direction. A look mismatch is not grounds to re-roll. Present the decision visually: write an options payload with the assigned direction leading, its raised lines included; the pick card when one exists; the dealt challengers as alternates with their QUALITY BAR cards, verdicts, and kept lines; re-roll with its safer and bolder registers; steer; canon enabled; and `buildPath` carrying the recorded default with `toggle: true` whenever image generation exists (details in the build-path paragraph below). A degraded roll with no challengers still uses the page, as a single text-only card with re-roll. Give every card the same anatomy: thesis, palette, materials, first viewport, honest risk, and the challengers' case lines (`--schema` prints the exact shape); the page renders identity from these fields, demotes declined challengers to their row on its own, and a challenger's catalog image rides as labeled inspiration, never the promise of the build. Author `canonCard` too: the category standard as one honest card, same anatomy; the page keeps it subordinate, and the counterweights still bind you. Run `node .pi/skills/impeccable/scripts/serve-question.mjs --start --payload <file>` (`--schema` first for the payload shape). It daemonizes, prints the page URL and a key, and exits; open that URL for the user, in-app browser first, then the system opener, then showing the URL. Collect the choice with `--wait --key <key>`, repeating while it exits 3; the ANSWER prints as JSON. An ANSWER of `{"optionId":"reroll"}` keeps the server alive and the page open on a loading hand: rerun concept-seed with the same `--scope` and `--mode` plus `--from <seed-key> --reroll <n>` (1 on the first re-roll, counting up), build the next payload, deliver it with `--update --key <same key> --payload <file>`, then return to `--wait` on that key. Never `--start` a second server or fall back to chat here: either strands the open tab on a hand that never arrives. Exit 4 means the page closed unanswered: re-present once through the structured question tool, and with no answer there either, proceed unattended with the assigned direction and state the assumptions. A harness that can leave a shell blocked in the background may run the script without `--start` and let it auto-open and block. Never predict the fallback: run the script, and only exit code 2 from starting it routes the decision to the structured tool; that exit is the fallback, never an error to retry.

 When image generation exists, every card also declares a `comp` path under `.impeccable/mocks/decision/`, the canon card included. Where the harness sandboxes its shell, start the page through the least-sandboxed command path it offers: a sandboxed shell cannot bind the board's port, and the first-attempt failure costs a retry every session. Serve the page first, then produce the comps; the page shimmer-waits per slot and the user may answer before they land. Each card's image is that direction's north-star comp at full fidelity under [visualize.md](visualize.md)'s comp discipline: the requested surface's first viewport, structure-led prompt, real product name and real content, no invented commercial claims, in that card's own palette, type character, and material world, committed all the way. Generation takes the same time at any fidelity, so an unfinished draft pays comp cost for draft quality; fairness between cards is equal fidelity in each card's own grammar, one surface, one aspect, never shared unfinishedness. The frame's aspect is the surface's own: portrait at device viewport for a native app or mobile-first surface, landscape for desktop web; the decision page adapts to either, and a phone screen comped landscape is a broken frame, not a neutral default. Produce in reading order, the assigned card, then the pick, then the full-card hand, then canon, each file written with its prompt sidecar the moment it is done, so a re-roll's spend front-loads onto the cards read first; declined challengers get no comp, their catalog thumb is their face. With parallel subagents, fan out one agent per card: each spawn is the shipped asset producer with a single-comp packet, that card's fields, PRODUCT.md, the shared frame, and the card's declared path, up to four in flight. Regenerate inline any slot still empty when its agent returns; drop without ceremony any slot still empty when the user answers. No other supervision is owed. Without parallel subagents, generate in the main thread after serving, same order, and let the harness's own generation display carry the progress; the wait for the answer follows the last file. The chosen card's comp is not spent by the choice: comp-led, it enters the comp round as compositional option one; code-led, it returns at the finish review as the critique reference, what the image dared that the build did not. Unchosen comps stay in `.impeccable/mocks/decision/` as the round's spent hand; they carry no approval and imply none. With no image generation, cards carry their identity in palette chips and facts, and that page is complete, not a lesser version; the page then also demotes every challenger's catalog art to a labeled thumbnail on its own, because salience must encode the verdict, never the accident of which cards have images.

@@ -29,10 +29,10 @@ Use the feature yourself at the surface's representative sizes: desktop and mobi
 If a prior critique exists, use it as one input:

 ```bash
-node .pi/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>"
+node .pi/skills/impeccable/scripts/critique-storage.mjs latest "<resolved target>" --json
 ```

-Exit 0 returns the latest snapshot; incorporate relevant P0/P1 findings and name the snapshot read. Exit 2 means none exists. Perform an independent pass either way.
+Exit 0 returns JSON with the latest snapshot's `body` and an exact `snapshot_file` identity. Retain `snapshot_file` until the end of the pass. For a local file target, the helper compares the file's exact current content fingerprint with the fingerprint captured by critique. Unchanged staged, unstaged, or untracked content remains current; any byte change, deletion, or replacement with a non-file closes the backlog it identified while preserving its trend history and exits 2. A URL target has no local fingerprint and remains current until explicitly closed. When current, incorporate relevant P0/P1 findings from `body` and name the snapshot read. Exit 2 means none exists or the target changed. Perform an independent pass either way.

 ## 3. Triage

@@ -95,3 +95,11 @@ Walk the complete path again with mouse, keyboard, and touch where applicable. C
 Follow the quality guidance supplied by `context.mjs` and hooks, then run any other relevant QA commands. Context requests a manual scan only when no automatic detector is active; never add another detector pass. Fix real defects and document only narrow intentional exceptions. A clean scan does not replace visual judgment.

 Finish with a source diff: remove accidental churn, orphaned code, redundant values, and temporary artifacts. Ship only when the feature is functionally complete and consistently finished across the path.
+
+When this pass clears every Priority Issue it took from a snapshot, close that snapshot:
+
+```bash
+node .pi/skills/impeccable/scripts/critique-storage.mjs close "<resolved target>" "<snapshot_file returned by latest>"
+```
+
+This closes only the snapshot this pass actually processed; if a newer critique landed meanwhile, its backlog stays live. Do not close when no snapshot was read, when `snapshot_file` was not retained, or when Priority Issues remain.
@@ -8,7 +8,7 @@ Reason over the signals; there is no score to obey:

 - `setup.hasDesign` false while `setup.hasCode` true → `document` (capture the visual system).
 - `critique.latest` is `null` → the project has never been critiqued; for a set-up project with a real surface, offering `/impeccable critique <surface>` is a strong default.
- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog), or re-run `critique` if the snapshot looks stale.
+- `critique.latest` with a low `score` or non-zero `p0` / `p1` → `polish` (it reads that snapshot as its backlog and closes it when stale or cleared).
 - `git.changedFiles` pointing at one surface → scope `audit` or `polish` to those files specifically, naming them.
 - `devServer.running` true → `live` is available for in-browser iteration; if false, don't lead with `live`. **`live` and the bundled `detect.mjs` are web-only.** If `setup.platform` is `ios`, `android`, or `adaptive`, don't lead with either; the browser overlay and the HTML rule engine don't apply to native app code.
 - Otherwise group by intent (build new / improve what's there / iterate visually), tailored to the current surface and `setup.platform`.
--- a/Show More
+++ b/Show More