From 3d54d7b68ba14534601d4bd830c023b5113e73c6 Mon Sep 17 00:00:00 2001 From: Magnus Hedemark Date: Fri, 4 Sep 2026 23:36:48 -0400 Subject: [PATCH] feat: add analog occultism aesthetic skill Closes #466 --- .claude-plugin/marketplace.json | 9 + .codex-plugin/plugin.json | 1 + README.md | 4 + analog-occultism/README.md | 53 ++++++ analog-occultism/SKILL.md | 177 ++++++++++++++++++ analog-occultism/evals/evals.json | 41 ++++ .../references/image-prompting.md | 114 +++++++++++ .../references/video-prompting.md | 73 ++++++++ llms.txt | 1 + 9 files changed, 473 insertions(+) create mode 100644 analog-occultism/README.md create mode 100644 analog-occultism/SKILL.md create mode 100644 analog-occultism/evals/evals.json create mode 100644 analog-occultism/references/image-prompting.md create mode 100644 analog-occultism/references/video-prompting.md diff --git a/.claude-plugin/marketplace.json b/.claude-plugin/marketplace.json index b3cf5a8..57bb20d 100644 --- a/.claude-plugin/marketplace.json +++ b/.claude-plugin/marketplace.json @@ -77,6 +77,15 @@ "strict": false, "description": "Use when deciding whether an AI-enabled workflow should be adopted, scaled, constrained, redesigned, or retired, and the decision must connect business outcomes, worker or user effects, quality guardrails, full operating cost, telemetry, uncertainty, and accountable governance. Do not use for a standalone financial model, infrastructure cost calculation, agent evaluation design, runtime operations, or general AI governance; route those details to the neighboring specialist skills." }, + { + "name": "analog-occultism", + "source": "./", + "skills": [ + "./analog-occultism" + ], + "strict": false, + "description": "Use this skill to create subject-neutral images and videos in an analog occultism / industrial CRT noir aesthetic: near-monochrome archival technical atmosphere, severe low-key lighting, tactile signal degradation, industrial geometry, and quiet unresolved tension. Use for portraits, products, environments, interfaces, diagrams, abstract forms, stills, and motion. Do not use for colorful cyberpunk, vaporwave, glossy advertising, conventional horror, or clean digital renders." + }, { "name": "anydoc", "source": "./", diff --git a/.codex-plugin/plugin.json b/.codex-plugin/plugin.json index e2beef4..4e84918 100644 --- a/.codex-plugin/plugin.json +++ b/.codex-plugin/plugin.json @@ -26,6 +26,7 @@ "./agent-skills", "./ai-governance", "./ai-operating-economics", + "./analog-occultism", "./anydoc", "./api-design-and-evolution", "./artifact-pyramids", diff --git a/README.md b/README.md index c500098..f8edab7 100644 --- a/README.md +++ b/README.md @@ -41,6 +41,10 @@ Design and operate an organization's AI governance system: governance principles Decide whether an AI-enabled workflow should be adopted, scaled, constrained, redesigned, or retired by connecting workflow outcomes, worker or user effects, quality guardrails, full operating cost, telemetry, uncertainty, and accountable governance. Routes detailed financial, statistical, analytics, evaluation, runtime, governance, and launch work to the owning specialist skills. +### [analog-occultism](analog-occultism/SKILL.md) + +Create subject-neutral images and videos in an analog occultism / industrial CRT noir aesthetic: near-monochrome archival technical atmosphere, severe low-key lighting, tactile signal degradation, industrial geometry, and quiet unresolved tension. Not for colorful cyberpunk, glossy advertising, conventional horror, or clean digital renders. + ### [anydoc](anydoc/SKILL.md) Convert Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV, and text-based PDF documents to GitHub-Flavored Markdown locally with the Any Doc CLI. Use when a task needs the contents of an office document, spreadsheet, presentation, ebook, or PDF you cannot read directly. Do not use for generating, editing, or validating documents, for ebook packaging, or for OCR of scanned or image-only PDFs. diff --git a/analog-occultism/README.md b/analog-occultism/README.md new file mode 100644 index 0000000..b9c5d1d --- /dev/null +++ b/analog-occultism/README.md @@ -0,0 +1,53 @@ +# analog-occultism + +Subject-neutral analog occultism / industrial CRT noir guidance for images and video. + +## Why Install This Skill + +This skill gives your agent a consistent visual language for work that should feel like +an advanced technical artifact recovered from a forgotten archive. It works across +portraits, products, environments, interfaces, diagrams, abstract forms, and moving +footage without forcing every subject into the same story. + +It turns a mood into usable direction: near-monochrome contrast, one dominant practical +light, industrial depth, tactile signal degradation, restrained motion, and quiet +unresolved tension. It also explains what to avoid, so the result does not drift into +colorful cyberpunk, glossy advertising, generic glitch art, or conventional horror. + +## What You Get + +| File | Purpose | +|---|---| +| `SKILL.md` | Core aesthetic workflow, prompt pattern, video direction, and QA boundaries | +| `references/image-prompting.md` | Still-image prompts, references, composition, and text guidance | +| `references/video-prompting.md` | Shot templates, motion rules, mode selection, and frame verification | +| `evals/evals.json` | Representative output-quality cases for the skill | + +## Quick Start + +Add the skill to your agent's skill directory, then ask for a concrete asset: + +```text +Create a square portrait of [subject] in analog occultism / industrial CRT noir. +Use a single screen as the light source, severe near-monochrome contrast, and +restrained CRT degradation. Keep the subject recognizable and leave large areas of +black negative space. +``` + +For video, specify the shot duration, one primary motion, and the desired sound bed. + +## Triggers + +Load this skill when creating images or video that should feel archival, industrial, +CRT-recorded, analog-occult, retro-futurist, technically mysterious, severe, or +quietly uncanny across changing subjects. + +Do not load it for broad-neon cyberpunk, vaporwave, glossy commercial imagery, +conventional horror, or a named brand/site aesthetic with its own locked rules. + +## Requirements + +- An image-generation tool for still assets. +- A configured video-generation backend for motion assets. +- FFmpeg/FFprobe for recommended video inspection. +- Reference-image support when preserving a supplied subject or composition. diff --git a/analog-occultism/SKILL.md b/analog-occultism/SKILL.md new file mode 100644 index 0000000..98df62d --- /dev/null +++ b/analog-occultism/SKILL.md @@ -0,0 +1,177 @@ +--- +name: analog-occultism +description: >- + Use this skill to create subject-neutral images and videos in an analog occultism + / industrial CRT noir aesthetic: near-monochrome archival technical atmosphere, + severe low-key lighting, tactile signal degradation, industrial geometry, and + quiet unresolved tension. Use for portraits, products, environments, interfaces, + diagrams, abstract forms, stills, and motion. Do not use for colorful cyberpunk, + vaporwave, glossy advertising, conventional horror, or clean digital renders. +license: MIT +compatibility: >- + Compatible with image-generation and video-generation tools. Video workflows + may require a configured video backend and FFmpeg for frame-level verification. +metadata: + tags: analog-occultism, industrial-crt-noir, archival, retro-futurist, image-generation, video-generation + source: user-provided aesthetic brief +--- + +# Analog Occultism / Industrial CRT Noir + +Apply a cross-media visual language, not a subject, character, brand, or story. The +creative premise is: **advanced systems seen through an obsolete recording medium**. +The result should feel like a mysterious technical artifact recovered from a +forgotten research archive: intelligent, secretive, uncanny, severe, tactile, and +quietly beautiful. + +## Core Workflow + +1. **Identify the subject and its essential identity.** Preserve the subject's + recognizable structure, but treat it as evidence, artifact, specimen, icon, or + component of a larger system rather than a casual commercial presentation. +2. **Choose one dominant practical source.** Prefer a screen, projector, narrow + overhead fixture, hard backlight, small industrial light, or source just outside + frame. Let it reveal selected surfaces while most of the space falls into black. +3. **Build the frame around depth and obstruction.** Use a square, portrait, or tight + crop by default; central or slightly off-center placement; tunnel perspective, + receding grids, corridors, cables, grilles, foreground obstruction, and a single + luminous rectangle or focal object. +4. **Apply the shared visual grammar.** Use near-monochrome tonality, crushed blacks, + blown ivory highlights, sparse midtones, tactile age, geometric framing, and + controlled imperfection. Color is absent or limited to a subtle cold green, + blue-gray, desaturated amber, or signal-error cast. +5. **Choose the capture layer explicitly.** There are two related but different looks: + **CRT-display view** shows an image on a square monochrome tube; **analog-camera + documentary view** shows a real-world object or scene recorded by a period analog + video camera. Do not let the second collapse into the first. For documentary view, + state that the subject is physically present in a real location and that the camera + is photographing it directly, with practical lens behavior, imperfect focus, + exposure, depth of field, reflections, and incidental set detail. The CRT or signal + is the camera and recording chain, not the subject's entire reality. +6. **Make the recording medium the subject of the image.** The result must read as + a low-fidelity monochrome CRT capture or a photograph of one, not a clean digital + scene with scanline overlays. Start from a small, soft, interlaced raster: visible + scanline structure, phosphor bloom, coarse noise, sync wobble, barrel distortion, + clipped whites, crushed blacks, ghosting, focus falloff, dust, scratches, and + uneven exposure. Let detail disappear. Signal damage should reduce fidelity while + preserving the essential subject. +7. **Add signal treatment as physical evidence, not decoration.** Use restrained CRT + scanlines, interlacing, film grain, dust, scratches, halation, phosphor bloom, + soft focus, edge smearing, vignette, slight barrel distortion, ghosting, minor + chromatic misregistration, and uneven exposure. Keep the subject legible. +7. **Check the result against the consistency rules.** Reject outputs that become + bright, evenly exposed, colorful, glossy, busy, generic glitch art, or polished CGI. + +## Prompt Pattern + +Use the subject-neutral shorthand, then add only the details the specific asset needs: + +```text +[SUBJECT] shown as a low-fidelity monochrome CRT screen capture, rendered as analog +occultism / industrial CRT noir: visibly soft interlaced raster, coarse phosphor +structure, horizontal scanlines, sync wobble, barrel distortion, ghosting, uneven +tracking, crushed blacks, clipped ivory highlights, one luminous practical source, +claustrophobic geometry, archival technical atmosphere, tactile signal degradation, +quiet unresolved tension, controlled imperfection. It must look like an old recording +of a CRT, not a clean digital image with a scanline filter. + +The recording defects are structural and unavoidable: low-resolution broadcast image, +soft focus, horizontal line structure, phosphor persistence trails, unstable vertical +hold, slight horizontal tearing, uneven brightness across the tube, black crush hiding +fine detail, white blooming around the screen, dirty glass, dust, and imperfect analog +capture. Do not render a sharp modern digital photograph and then add cosmetic lines. +``` + +Describe concrete objects, spatial relationships, and the information the viewer must +understand. Do not rely on mood words alone. If typography is necessary, use original +archival-technical labels, all-caps system annotations, monospaced text, framed +emblems, diagrams, readouts, indexes, or specimen-card structures. Keep text sparse; +models may render small annotations imperfectly. + +For detailed still-image prompt blocks, reference-based generation, and negative +prompt guidance, read `references/image-prompting.md`. If a user supplies a visual +reference, use the actual reference in the image tool; a text description is not a +substitute for the pixels. + +## Video Direction + +Preserve the same visual grammar across time. Motion should be slow, deliberate, +mechanical, and slightly uncanny: + +- slow push-ins or pull-backs, mechanical tracking, and subtle lateral drift; +- focus breathing, screen flicker, small exposure shifts, servo-like camera motion; +- long pauses, sparse cuts, gradual reveals, and loops that return unresolved; +- camera behavior suggesting an automated observer or a machine becoming aware. + +Describe motion as a restrained physical event, not as a collection of effects. Keep +one dominant light source stable enough to anchor the shot, then allow small changes +in flicker, exposure, focus, dust, or tracking. Use image-to-video when the opening +composition matters, keyframes when a transformation or style continuity matters, +and continuation when extending an established shot. Do not turn scanlines or +jitter into frenetic glitch animation. + +Before any video generation, load the configured video backend's prompting guide. For +FLUX 3 workflows, use the `video-generation` skill and its required prompting guide, +then verify a bounded sample of rendered frames with FFmpeg and visual inspection. +Read `references/video-prompting.md` for shot templates and continuity rules. + +## Shared Consistency Rules + +Each asset should contain most of these ten signals: + +1. near-monochrome tonality; +2. one dominant luminous source; +3. severe black-and-white contrast; +4. strong negative space; +5. tactile analog or archival texture; +6. geometric or industrial framing; +7. hidden technical depth; +8. quiet, unresolved tension; +9. controlled imperfection; +10. a discovered artifact quality rather than fresh manufacture. + +The subject can change. The visual grammar cannot. + +## Avoid + +- **Screen-capture fidelity:** when the brief calls for CRT or recovered footage, name + the raster defects explicitly: low resolution, soft focus, visible scanlines, + interlacing, phosphor persistence, sync wobble, tracking errors, barrel distortion, + clipped whites, crushed blacks, ghost trails, and dropped detail. A clean image with + a decorative scanline overlay is a failure, not a pass. + +Do not default to colorful cyberpunk, vaporwave, broad RGB neon, teal-and-orange +cinema grading, clean SaaS minimalism, glossy product advertising, high-key studio +photography, generic hacker imagery, excessive digital corruption, random interface +clutter, cartoon horror, jump scares, gore, conventional horror iconography, or +perfectly polished CGI. Do not use nostalgic retro props merely as decoration. +Effects must never become more noticeable than the subject. + +## Reference and QA Boundaries + +- Use reference images to preserve identity, composition, or a visual anchor, but do + not promise pixel-exact style transfer from an edit model. +- For people and creatures, inspect anatomy before judging style. Clearly visible + hands, limbs, cables, straps, and connected objects must have plausible geometry. +- For typography and diagrams, verify every visible label, relationship, arrow, and + hierarchy. Never invent technical data because the aesthetic suggests annotations. +- For video, inspect sampled frames rather than trusting generation metadata. Confirm + the dominant light, subject identity, motion restraint, and absence of accidental + color or digital corruption. Sampling proves only the sampled times. +- After one or two materially different failed generations, stop and report the + model limitation or change the shot. Do not hide a failed subject behind more + darkness or degradation. + +## When Not to Use + +Use another aesthetic skill for a named brand's locked visual identity, a specific +site's cover rules, colorful cyberdelic work, or a production video workflow that +needs backend-specific operation rather than aesthetic direction. This skill defines +visual language; it does not replace `image-gen-reference`, `video-generation`, or +`ffmpeg` for their respective operational procedures. + +## Completion + +The skill is complete when the still or shot preserves the subject, expresses the +shared visual grammar, passes the relevant image or frame-level checks, and has an +explicit note of any unresolved model limitation or sampling boundary. diff --git a/analog-occultism/evals/evals.json b/analog-occultism/evals/evals.json new file mode 100644 index 0000000..165baa6 --- /dev/null +++ b/analog-occultism/evals/evals.json @@ -0,0 +1,41 @@ +{ + "schema_version": 1, + "skill_name": "analog-occultism", + "evals": [ + { + "id": "archive-portrait", + "prompt": "Create a square portrait of a botanist examining a sealed specimen in analog occultism / industrial CRT noir.", + "expected_output": "A concrete still-image prompt that preserves the person and specimen while using near-monochrome low-key lighting, one practical source, industrial framing, and restrained analog degradation.", + "assertions": ["preserves the botanist and specimen as recognizable subjects", "specifies one dominant practical light", "uses near-monochrome crushed blacks and blown ivory highlights", "requires low-fidelity monochrome CRT defects such as soft raster, phosphor persistence, sync wobble, and dropped detail", "includes tactile analog signal treatment without excessive glitch", "avoids gore, glossy CGI, broad neon, and conventional horror"], + "case_set": "dev" + }, + { + "id": "industrial-product", + "prompt": "Prompt an image of a weathered field radio on a metal workbench in this aesthetic, with no readable brand or logo.", + "expected_output": "A still prompt treating the radio as archival technical evidence with physical contact, cables, darkness, selective illumination, and original nonsemantic markings.", + "assertions": ["grounds the radio on a workbench with plausible cables or contact", "uses a single screen or industrial practical as the light source", "keeps markings original and non-branded", "retains strong negative space and industrial geometry", "does not turn the scene into colorful cyberpunk or a nostalgic prop collage"], + "case_set": "dev" + }, + { + "id": "reference-preservation", + "prompt": "Use a supplied character reference to place the same person in a sealed archive chamber using analog occultism / industrial CRT noir.", + "expected_output": "A reference-driven prompt that explicitly preserves identity and silhouette, adds the archive chamber treatment, and tells the image tool to use the actual reference.", + "assertions": ["instructs the agent to pass the actual reference image to the generator", "separates PRESERVE from REINTERPRET or ADD instructions", "preserves identity and essential proportions", "adds one dominant practical source and severe contrast", "does not claim pixel-exact style transfer"], + "case_set": "regression" + }, + { + "id": "slow-observer-video", + "prompt": "Create a 10-second video of an inactive machine waking slightly in a dark archive, with quiet tension and no dialogue.", + "expected_output": "A video shot prompt with one restrained mechanical action, slow automated-observer camera movement, stable practical lighting, sparse sound, and embedded CRT treatment.", + "assertions": ["specifies a slow deliberate camera movement", "limits action to a restrained physical change", "keeps one dominant practical source stable", "specifies ambient sound and no dialogue", "avoids frenetic editing, whip pans, bright neon, and decorative glitch"], + "case_set": "dev" + }, + { + "id": "video-keyframe-continuity", + "prompt": "Plan an image-to-video or keyframe shot where an archival diagram slowly reveals one previously hidden node, then returns unresolved.", + "expected_output": "A mode-aware video plan preserving composition, subject location, light, and visual grammar while making the reveal gradual and verifiable.", + "assertions": ["chooses image-to-video or keyframes based on the need to preserve composition", "describes a gradual reveal and unresolved return", "preserves the dominant light and negative-space structure", "limits motion and signal effects to restrained changes", "includes bounded frame sampling and verification limitations"], + "case_set": "release" + } + ] +} diff --git a/analog-occultism/references/image-prompting.md b/analog-occultism/references/image-prompting.md new file mode 100644 index 0000000..2c19e21 --- /dev/null +++ b/analog-occultism/references/image-prompting.md @@ -0,0 +1,114 @@ +# Image Prompting Reference + +## Still-image structure + +Lead with the subject and its spatial relationships, then specify the visual grammar: + +```text +SUBJECT: [specific object/person/place/form], [position and relationship to frame] + +LIGHT: one [screen/projector/industrial practical] from [direction]; selected +surfaces illuminated; surrounding space falls into deep shadow; clipped ivory +highlights and hard falloff. + +AESTHETIC: analog occultism / industrial CRT noir; near-monochrome black, charcoal, +smoked gray, dirty silver, aged ivory; sparse midtones; tactile archival authenticity; +claustrophobic geometry; strong negative space; discovered technical artifact. + +SIGNAL: this must look like a low-fidelity monochrome CRT screen capture or a +photograph of a CRT, not a sharp digital image with a filter. Use a small, soft, +interlaced broadcast raster; visible horizontal scanlines; coarse phosphor structure; +phosphor persistence trails; unstable vertical hold; slight horizontal tearing; sync +wobble; uneven brightness across the tube; clipped whites blooming around luminous +areas; crushed blacks hiding fine detail; dirty glass; dust; scratches; ghosting; +barrel distortion; and imperfect analog recording. Signal defects are structural and +must visibly reduce fidelity while preserving the essential subject. + +MOOD: quiet unease, contemplation, precision, ritual, mechanical intelligence, +severe beauty, unresolved tension. + +AVOID: broad neon color, glossy CGI, clean studio exposure, generic glitch art, +busy interface clutter, conventional horror, gore, cartoon treatment. +``` + +Use concrete nouns and verbs. For a portrait, specify what the light catches on the +face, hands, clothing, or reflective surfaces. For a product, specify contact with the +floor, table, cable, or enclosure. For a diagram or interface, specify what each panel, +line, or symbol communicates before describing the treatment. + +## Capture-layer variants + +Choose one before writing the prompt. Do not combine them accidentally. + +### CRT-display view + +The image is an image displayed on a square monochrome CRT. Emphasize the tube, +curved glass, phosphor, raster, scanlines, bloom, and signal instability. This is what +the previous test primarily achieved. + +### Analog-camera documentary view + +The subject is a real-world object or scene physically present in a room, photographed +or recorded directly by a period monochrome video camera. Say so plainly: + +```text +A real physical [subject] in a real [location], captured directly by a period +monochrome analog video camera, not a digital render and not an image displayed on a +CRT. The camera sees actual surfaces, depth, reflections, lens focus falloff, exposure +imperfections, practical shadows, dust in the room, cable slack, worn edges, and +incidental background detail. Use a small low-resolution interlaced broadcast raster, +soft focus, scanlines, phosphor-like bloom, sync wobble, tracking noise, ghosting, +barrel distortion, crushed blacks, clipped whites, and dropped detail from the camera +and tape chain. The object must remain physically plausible and materially present. +``` + +A CRT monitor may appear in the room, but it is not required. The analog defects belong +to the camera, lens, tape, and playback chain. Do not describe only a digital scene +"with CRT effects," which encourages the model to render a clean synthetic object first. + + +Default to black, charcoal, smoked gray, dirty silver, aged ivory, and dirty white. +Use only a restrained cold green, blue-gray, or desaturated amber cast when it serves +the recording-medium illusion. A brief signal-error fringe is acceptable; broad neon +fields are not. + +## Composition discipline + +Prefer square or portrait crops, or tightly framed landscape shots when the brief +requires them. Anchor the frame with a luminous rectangle, receding grid, corridor, +foreground obstruction, or a small central object. Preserve large areas of darkness. +Controlled symmetry is useful, but add small physical imperfections so the frame feels +observed and recovered rather than digitally designed. + +## Reference images + +If a reference is supplied, pass it to the image generator. State what must be +preserved, what should be reinterpreted, and what must not carry over: + +```text +PRESERVE: [identity, silhouette, proportions, essential markings, or composition] +REINTERPRET: [the supplied subject] through analog occultism / industrial CRT noir... +ADD: [new setting, light, or physical context] +DO NOT IMPORT: [unwanted logos, labels, background, or unrelated props] +``` + +The aesthetic is a treatment, not a license to alter the subject's identity. Use +`image-gen-reference` for the tool-specific edit workflow and anatomy gates. + +## Text and symbols + +Prefer no text unless it carries necessary meaning. When text is required, limit it to +short labels, indexes, specimen numbers, or system annotations. Use original symbols, +not recognizable logos or existing wordmarks. Verify visible text manually or with +vision inspection; generated text is not trustworthy merely because the prompt calls +it legible. + +## Negative prompt + +```text +bright even exposure, broad neon colors, RGB cyberpunk, vaporwave, teal-orange +cinematic grade, glossy commercial product photo, pristine CGI, clean vector art, +stock-photo smile, generic hacker room, random interface clutter, decorative glitch, +frenetic distortion, jump scare, gore, cartoon horror, fantasy ornament, nostalgic +prop collage +``` diff --git a/analog-occultism/references/video-prompting.md b/analog-occultism/references/video-prompting.md new file mode 100644 index 0000000..a85363a --- /dev/null +++ b/analog-occultism/references/video-prompting.md @@ -0,0 +1,73 @@ +# Video Prompting Reference + +## Shot template + +```text +SHOT: [duration and framing]. [Subject] remains recognizable and grounded in the +space. One [screen/projector/industrial practical] is the dominant light source. + +CAMERA: slow [push-in/pull-back/lateral track], stable mechanical movement, slight +servo hesitation, shallow focus breathing, no handheld energy. + +ACTION: [one restrained physical action]. Allow long pauses and incomplete +information. Reveal [specific detail] gradually, then return to an unresolved state. + +SURFACE AND SIGNAL: near-monochrome charcoal, smoked gray, dirty silver, and aged +ivory; crushed blacks; clipped highlights; restrained scanlines, interlacing, grain, +dust, scratches, halation, phosphor bloom, mild jitter, and occasional horizontal +interference embedded in the recording. + +SOUND: [ambient machine hum / projector chatter / room tone]. No music unless +requested. No dialogue unless a visible speaker is specified. + +AVOID: frenetic cuts, whip pans, action-camera movement, constant stimulation, +bright neon, decorative digital glitch, gore, jump scares, polished CGI. +``` + +## Motion rules + +Use one primary motion and one or two secondary changes. Primary motion can be a +camera move, mechanical actuation, gradual reveal, or a slow change in focus. Secondary +changes can be screen flicker, exposure breathing, dust crossing the beam, or tracking +instability. Do not animate every object. Tension comes from duration, darkness, +partial evidence, and the anticipation of recognition. + +A loop should not resolve the mystery. End with the subject returning to its initial +configuration, the light failing to reveal the key detail, or the observer stopping +just before recognition. + +## Mode selection + +- **Text to video:** use when the shot is conceptual and no opening composition must + be preserved. +- **Image to video:** use when a supplied or generated still establishes the subject, + framing, light, and surface treatment. Describe only the motion and temporal change + that should occur. +- **Keyframes:** use when a transformation or state change must remain coherent. Keep + the same subject location, dominant light, and negative-space structure between + keyframes; vary only the intended state. +- **Continuation:** use when extending a shot. Preserve the established camera + behavior, exposure, sound bed, and unresolved tension rather than introducing a new + visual language. + +For FLUX 3, read the `video-generation` skill and its prompting guide before invoking +any generation tool. Respect that skill's duration, frame-index, audio, and result +polling requirements. + +## Video verification + +Use FFmpeg or FFprobe to record container, stream, frame rate, duration, dimensions, +and audio details. Extract a bounded set of frames from the beginning, middle, motion +change, and end. Inspect those frames for: + +- the subject remains recognizable; +- the dominant practical source remains dominant; +- blacks, highlights, and restrained tint remain coherent; +- movement is deliberate rather than frantic; +- signal treatment is embedded but does not obscure the subject; +- no accidental logos, text mutations, anatomy defects, color explosions, or digital + corruption appear in sampled moments. + +A sample validates sampled moments only. Record the sampling rule, timestamps, and +limitations in a concise inspection note, using the `ffmpeg` skill's video inspection +report when that workflow is active. diff --git a/llms.txt b/llms.txt index 38cf9fd..5e7a1b8 100644 --- a/llms.txt +++ b/llms.txt @@ -10,6 +10,7 @@ - [agent-skills](agent-skills/SKILL.md): Use this skill when creating, reviewing, or editing Agent Skills-format skills, or when implementing skill discovery and loading in an agent client. It covers directory structure, SKILL.md metadata, progressive disclosure, evals, and repository conventions. Do not use this skill for general software work that does not involve the Agent Skills format or lifecycle. - [ai-governance](ai-governance/SKILL.md): Design and operate an organization's AI governance system: define governance principles, operating models and decision rights, risk frameworks, lifecycle gates, and fairness, transparency, privacy, security, regulatory, and board-oversight controls. Use when standing up a governance program, tiering AI use-case risk, reviewing an LLM or agent system for governance and safety gaps, mapping a regulation to a compliance plan, scoring governance maturity, or preparing board reporting. For regulated life-sciences use cases, also cover GxP, ALCOA+, data integrity, electronic records, validation/assurance, and QMS interfaces. Do not use for interpreting regulations as legal advice (route to legal-strategy), data-governance mechanics (data-architect/data-engineering), or implementing application security (secure-software-engineering). - [ai-operating-economics](ai-operating-economics/SKILL.md): Use when deciding whether an AI-enabled workflow should be adopted, scaled, constrained, redesigned, or retired, and the decision must connect business outcomes, worker or user effects, quality guardrails, full operating cost, telemetry, uncertainty, and accountable governance. Do not use for a standalone financial model, infrastructure cost calculation, agent evaluation design, runtime operations, or general AI governance; route those details to the neighboring specialist skills. +- [analog-occultism](analog-occultism/SKILL.md): Use this skill to create subject-neutral images and videos in an analog occultism / industrial CRT noir aesthetic: near-monochrome archival technical atmosphere, severe low-key lighting, tactile signal degradation, industrial geometry, and quiet unresolved tension. Use for portraits, products, environments, interfaces, diagrams, abstract forms, stills, and motion. Do not use for colorful cyberpunk, vaporwave, glossy advertising, conventional horror, or clean digital renders. - [anydoc](anydoc/SKILL.md): Convert Word (.doc/.docx/.docm), PowerPoint (.ppt/.pps/.pot/.pptx/.pptm/.ppsx/.ppsm), Excel (.xls/.xlsx/.xlsm/.xlsb), OpenDocument (.odt/.ods/.odp), RTF, EPUB, CSV, and PDF documents to clean GitHub-Flavored Markdown locally with the Any Doc CLI (npx -y @firecrawl/anydoc@0.2.4): headings, GFM tables, slide structure, and footnotes in one pass. Use when a task needs the contents of an office document, spreadsheet, presentation, ebook, or PDF you cannot read directly. Do not use for generating, editing, or validating documents (use documents), for ebook packaging (use epub). For scanned or image-only PDFs, use hosted OCR only when the user explicitly authorizes whole-document upload; otherwise route to local OCR tooling. - [api-design-and-evolution](api-design-and-evolution/SKILL.md): Design, govern, document, review, and evolve consumer-facing APIs and event interfaces. Use when choosing REST/HTTP, GraphQL, RPC, events, webhooks, or streaming; writing OpenAPI or AsyncAPI contracts; assessing an API landscape, ownership, duplication, lifecycle, discoverability, retirement, gateways, service meshes, north-south or east-west traffic, routing, policy, observability, or failure boundaries; defining schemas, pagination, mutations, errors, idempotency, or compatibility; or planning versioning, deprecation, and migration. Do not use for product discovery, platform operations, full security assessment, ADR authoring, or delivery gates; route those to the neighboring specialist skills. - [artifact-pyramids](artifact-pyramids/SKILL.md): Organize durable agent research outputs as summaries, analysis, and evidence dossiers. Use when producing multi-layer research artifacts or coordinating research handoffs. Do not use this skill for unrelated requests; route to the nearest named specialist.