mirror of
https://github.com/magnus919/agent-skills.git
synced 2026-09-11 19:47:12 +03:00
Magnus HedemarkandGitHub
e10508b034
* feat(bmad): add BMad control-plane protocol skill New standalone methodology skill that lets any agent run the BMad method (Breakthrough Method of Agile AI-Driven Development) as a harness-agnostic control-plane protocol: five-field intent contracts, direct/bounded/initiative classification, review-as-triage, failure routing by layer, and autonomy gating with machine-readable spec status. - SKILL.md protocol core with progressive disclosure + When not to use - README.md human-facing install guide - 9 references: protocol, classification, spec, lifecycle, project-context, review-and-failure-routing, autonomy, party-mode, adoption - 4 templates: SPEC, INTENT, STORY, REVIEW - scripts/check-spec.py + 16 tests (stdlib, deterministic spec validation) - evals/evals.json: 9 output-quality cases - Routing seams from bmad to adjacent skills and back from spec-driven-development, product-shaping, implementation-planning, neckbeard - Catalog updates: root README, skill-triggers, marketplace/plugin/llms.txt Closes #399 * fix(bmad): address droid-review findings - check-spec.py: skip headings inside fenced/indented code blocks so a spec cannot PASS on section text that only appears in a code sample - check-spec.py: catch UnicodeDecodeError on non-UTF-8 files and report FAIL instead of crashing - STORY.md template: add created key for resumability/traceability parity - SPEC.md template: split in-progress and in-review status bullets - add 2 regression tests (heading-in-fence, non-UTF-8) * fix(bmad): address droid-review round 2 - check-spec.py: read specs with utf-8-sig so a UTF-8 BOM cannot silently disable the frontmatter status check - check-spec.py: handle standard YAML inline comments after status values (status: draft # pending review) without a false FAIL - references/protocol.md: make lifecycle phrasing consistent with lifecycle.md — four phases plus a learning closeout - add 2 regression tests (BOM, inline comment) * fix(bmad): tolerate trailing whitespace on frontmatter delimiters A spec whose --- delimiter lines carry trailing spaces or tabs would silently disable the status check and let an invalid status PASS. Relax the delimiter pattern and add a regression test. * fix(bmad): ignore inline comments in quoted status values * fix(bmad): tolerate leading blank lines before frontmatter * fix(bmad): fail closed on unparseable frontmatter, matching fence markers Address droid-review round 5 and 6 findings as a single closed class: - Fail closed when a file opens with a --- delimiter that cannot be parsed, so no whitespace/frontmatter permutation can silently disable the status check (previously: unparseable frontmatter was treated as 'no status' warning, letting an invalid status PASS). - Track fence opener markers in collect_headings so a mismatched fence no longer closes a code block early (false-PASS on missing sections) and an unclosed fence no longer swallows real headings. - Accept empty well-formed frontmatter (---\n---) and closing delimiters without a trailing newline. - STORY.md template: parent-spec points at the sibling SPEC.md. - README: status vocabulary is not a strict linear chain; blocked is a resumable routing signal. Whitespace/frontmatter mutation sweep: 9 formatting variants x valid/invalid status all verdict correctly; malformed delimiters fail closed. 29 tests.
7.3 KiB
7.3 KiB
name, description, license, metadata
| name | description | license | metadata | ||
|---|---|---|---|---|---|
| product-shaping | Use this skill to shape product or engineering work before committing time to it: set appetites instead of estimates, narrow raw ideas into bounded problems, sketch solutions at the right level of abstraction, de-risk rabbit holes, write pitches, bet with capped downside (circuit breaker), and govern builds with discovered scopes and scope hammering. Adapted from Basecamp's Shape Up and extended for human+AI-agent teams. Use when a raw idea, feature request, or "redesign X" grab-bag needs to become a bounded project before anyone builds; when planning how much work an idea is worth; or when delegated agent builds need budgets, kill criteria, and non-convergence rules. Do not use for discovering whether a problem is real (use product-discovery), for portfolio-level sequencing across quarters (product-roadmapping-and-portfolio), for formal specification after the bet is placed (spec-driven-development), or for task-level prioritization frameworks like RICE (product-methodology). | MIT |
|
Product Shaping
Pre-commitment methodology for product and engineering work, adapted from Ryan Singer's Shape Up (free edition at basecamp.com/shapeup), extended for teams whose builders include AI agents.
The loop: shape a raw idea into a bounded pitch → bet a fixed appetite on it → build by discovering scopes and hammering scope to fit the box → move on, letting post-ship feedback re-enter as raw ideas.
The core moves
- Set boundaries — choose the appetite ("how much is this worth?") and narrow the problem to one specific story. Kill grab-bags ("redesign X", "X 2.0").
- Find the elements — sketch the solution rough, solved, and bounded: breadboards for flows, fat-marker fidelity for visual problems, components-and-contracts for non-UI work.
- Patch rabbit holes — attack your own sketch; settle hard decisions now, declare out-of-bounds cases, cut what the appetite can't afford.
- Write the pitch — problem, appetite, solution, rabbit holes, no-gos.
- Bet — commit the box uninterrupted, downside capped. No finish, no extension by default: the circuit breaker routes failure back to shaping.
- Build — one integrated slice first, then discovered scopes tracked as uphill→downhill states; sequence scariest-first; compare down to baseline when deciding to stop.
- Move on — scope cuts are not quality cuts; new feedback needs shaping, not instant yes.
Reference files
Load only what the current step needs:
| Reference | Load when |
|---|---|
| references/principles.md | You need the why: appetite vs estimate, fixed-time-variable-scope, rough/solved/bounded, evidence boundaries, lineage |
| references/shaping.md | Shaping steps 1–4 in detail, including shaping non-UI/backend/infrastructure work |
| references/betting.md | Bets vs backlogs, circuit breaker mechanics, cycles as optional scaffolding, handling defects between bets |
| references/building.md | Hand-over-responsibility, one-piece-done, scope mapping, hill-state tracking, deciding when to stop |
| references/hybrid-adaptation.md | Any bet involving AI-agent builders: budget currencies, batched steering, verification cost inside scope, kill criteria for non-converging loops |
| references/anti-patterns.md | Before betting anything that matters — documented field failures and their mitigations |
Templates
| Template | Purpose |
|---|---|
| templates/PITCH.md | Fillable five-ingredient pitch document |
| templates/SCOPE_MAP.md | Fillable scope table with hill states, chowder list, and breaker check |
Entry points
| Situation | Start here |
|---|---|
| Raw idea or request arrived | references/shaping.md step 1 |
| Idea is validated but unbounded | references/shaping.md |
| Ready to write up the concept | templates/PITCH.md |
| Deciding what gets the next box | references/betting.md |
| Bet placed, starting the build | references/building.md |
| Builders are AI agents | references/hybrid-adaptation.md |
| Project keeps not finishing / loops won't converge | references/anti-patterns.md, then hybrid-adaptation kill criteria |
When not to use
- The problem itself isn't validated yet →
product-discovery. Shaping narrows validated problems; it does not investigate whether the problem is real. - The question is strategic (positioning, market entry, portfolio weight across
quarters) →
product-strategyorproduct-roadmapping-and-portfolio. This skill packages a single bet — one bounded commitment with an appetite and circuit breaker; roadmapping sequences many such bets across cycles with continue/pause/kill criteria. - The bet is already placed and the work needs a formal spec →
spec-driven- developmentconsumes shaped output when formal specification is warranted. - Comparing unrelated feature proposals by score →
product-methodology(RICE/ MoSCoW). Appetite replaces scoring inside this skill's scope; use one system, not both on the same decision. - An epic resists decomposition because nobody can define done → route BACK here: that is an unshaped project, and force-splitting it produces disconnected tasks.
- The work is small, routine, and fully understood — skip shaping overhead; just do it.
- The bet is placed and you need intent-to-delivery control — classification,
five-field intent contracts, autonomy gating, failure routing, and resumable status
across the run. Route to
bmad. Shaping ends at the bet; bmad carries the placed intent through bounded, inspectable, resumable agent work.
Related skills
product-discovery— upstream: validates the problem before narrowing beginsproduct-strategy,product-roadmapping-and-portfolio— strategic context above betsproduct-methodology— downstream consumer of a won bet (prioritization, spec drafting)spec-driven-development— optional formal specification of shaped output post-betimplementation-planning,subagent-driven-development— execution after the bet; decompose only downhill work, never pre-shred a pitchbmad— intent-to-delivery control-plane protocol once the bet is placed: classification, five-field intent contracts, autonomy gating, failure routing, resumable spec statuswork-tracking— where scope/hill state lives during the buildqa-methodology— edge-case QA as late-cycle level-up, not gate
Evidence note
Every practice claim originates from one company's account (Basecamp, 2019). Independent team records through 2026 show real adaptations and documented abandonments; read references/anti-patterns.md before betting anything that matters, and treat six-week cycles as tunable scaffolding rather than doctrine.