[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M72YB45DBYKD58QESJSE0C1P":3,"skill-same:01M72YB45DBYKD58QESJSE0C1P":65},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":35,"commitSha":36,"skillDirectory":37,"forksCount":38,"repositoryHtmlUrl":39,"repositoryHomepageUrl":40,"ownerHtmlUrl":41,"ownerBlogUrl":42,"ownerDescription":43,"updatedAt":44,"install":45,"resources":58},"01M72YB45DBYKD58QESJSE0C1P","clarify-prd",null,"Use to run an ambiguity sweep over a written PRD.md and close every under-specified point before glossary or architecture proceeds — so two engineers can't reasonably build different things from the same PRD. Triggers on \"\u002Fsdlc:clarify-prd {slug}\", \"clarify {slug}\", \"find ambiguities in {slug}\", \"is the PRD ready\", \"sharpen the PRD\", \"прояснити PRD\", \"знайди неоднозначності {slug}\", \"чи готова специфікація\". Re-reads the PRD, dispatches a clean-context devil's-advocate subagent to list where the PRD forks, then for each ambiguity runs AskUserQuestion to RESOLVE it (tighten §1\u002F§5\u002F§6 in place) or DEFER it (→ §8 Open questions with owner+due). Output: an updated docs\u002Ffeatures\u002F{slug}\u002FPRD.md with every ambiguity resolved or deferred — none dangling. Hard-refuse if PRD.md is missing.","genkovich\u002Fagentic-engineering-course-public","genkovich","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F14920737?v=4","01M72Y5BWEEZ8VT4DFW6AGGE8X","Kirill Sulimovsky","01M72Y5BXYRGV5PKCVCHEX81WY",15,"MIT","https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F6-sdlc\u002Fsdlc\u002Fplugin\u002Fskills\u002Fclarify-prd\u002FSKILL.md",[],[],"2026-11-10 22:39:42",false,"---\nname: clarify-prd\nmodel: opus\neffort: high\nagents: [devils-advocate]\ndescription: >-\n  Use to run an ambiguity sweep over a written PRD.md and close every under-specified\n  point before glossary or architecture proceeds — so two engineers can't reasonably build\n  different things from the same PRD. Triggers on \"\u002Fsdlc:clarify-prd {slug}\",\n  \"clarify {slug}\", \"find ambiguities in {slug}\", \"is the PRD ready\",\n  \"sharpen the PRD\", \"прояснити PRD\", \"знайди неоднозначності {slug}\", \"чи готова специфікація\".\n  Re-reads the PRD, dispatches a clean-context devil's-advocate subagent to list where the\n  PRD forks, then for each ambiguity runs AskUserQuestion to RESOLVE it (tighten §1\u002F§5\u002F§6\n  in place) or DEFER it (→ §8 Open questions with owner+due). Output: an updated\n  docs\u002Ffeatures\u002F{slug}\u002FPRD.md with every ambiguity resolved or deferred — none dangling.\n  Hard-refuse if PRD.md is missing.\ntriggers:\n  - \u002Fsdlc:clarify-prd\n  - \"clarify\"\n  - \"find ambiguities in\"\n  - \"is the PRD ready\"\n  - \"sharpen the PRD\"\nstage: \"03\"\n---\n\n# Skill: clarify-prd (SDLC stage 03 — PRD ambiguity sweep)\n\nAmbiguity sweep over a written `PRD.md`. It hunts the PRD for under-specified points — vague terms, unmeasured NFRs, AC missing error\u002Fauthz\u002Fedge behavior, unstated assumptions, conflicting requirements, undefined domain terms, missing actors, scope creep — then dispatches a **clean-context devil's-advocate subagent** that re-reads the PRD fresh and answers one question: *where would two engineers reasonably build different things from this?* Each ambiguity it surfaces is closed with the user: either **resolved** (the PRD is tightened in place) or **deferred** (a §8 Open-Questions row with owner + due). It exists so `fix-term` \u002F `architecture-design` never proceed on an ambiguous PRD.\n\nThis is a sweep, not a full authoring stage — it does **not** run the shared Socratic loop or the coherence critic. Its shared dependencies:\n→ [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md) (resolve\u002Fdefer question phrasing) · [`..\u002F_shared\u002Finterview-depth.md`](..\u002F_shared\u002Finterview-depth.md) (sweep aggressiveness + per-finding question volume) · [`..\u002F_shared\u002Fagent-roster.md`](..\u002F_shared\u002Fagent-roster.md) (dispatch discipline).\n\nDepth governs how aggressively the sweep hunts + the per-finding question volume. (Every surfaced ambiguity is still Resolved or Deferred at every level — that's a floor, not a dial.)\n\n## Owner\n\nPM + Tech Lead (the PRD's co-authors resolve their own ambiguities). PM owns vague-term \u002F scope \u002F missing-actor calls; Tech Lead owns unmeasured-NFR \u002F under-specified-AC \u002F conflicting-requirement calls.\n\n## Inputs\n\n- `\u003Cslug>` — same feature slug used by `write-prd`.\n- **Gate (hard-refuse if missing):** `docs\u002Ffeatures\u002F\u003Cslug>\u002FPRD.md`. Absent → STOP and point: «run `\u002Fsdlc:write-prd \u003Cslug>` first — clarify-prd sharpens an existing PRD, it does not write one».\n- (Optional) `docs\u002Ffeatures\u002F\u003Cslug>\u002FCONTEXT.md` — if present, its `## Glossary` is canonical; an \"undefined-term\" finding for a word already glossed there is a false positive (drop it).\n\n## Protocol\n\n1. **Gate + set interview depth.** `test -f docs\u002Ffeatures\u002F\u003Cslug>\u002FPRD.md` → missing = refuse with the pointer above. Read the PRD (and `CONTEXT.md` `## Glossary` if present, to suppress false \"undefined-term\" hits). **Then set the interview depth (the opening question):** read `interview_depth` from `.claude\u002Fsdlc.local.md` if present (else default medium), and — unless a `--depth=easy|medium|hard` arg was passed — ask ONE depth-selection `AskUserQuestion` phrased per [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md), with the saved\u002Fmedium value as the «(Recommended)» first option. The level tunes how adversarially the sweep + subagent hunt (easy: only build-divergence that changes behavior, with assumptions stated; medium: balanced; hard: adversarial, every fork surfaced) and the per-finding question volume → [`..\u002F_shared\u002Finterview-depth.md`](..\u002F_shared\u002Finterview-depth.md).\n\n2. **First-pass self-sweep.** Walk the PRD against the eight ambiguity classes in [`.\u002Freferences\u002Fambiguity-checks.md`](.\u002Freferences\u002Fambiguity-checks.md) (vague-term \u002F unmeasured-NFR \u002F under-specified-AC \u002F unstated-assumption \u002F conflicting-requirement \u002F undefined-term \u002F missing-actor \u002F scope-creep). Note candidate findings with a `§ref` each — do not edit yet.\n\n3. **Devil's-advocate subagent (the core mechanic).** Dispatch the `devils-advocate` agent — `subagent_type: \"sdlc:devils-advocate\"` (carries its own `model: opus` + `effort: high`; clean context — it never saw this conversation). Pass only the slug + the PRD path; it Reads `PRD.md` (and `CONTEXT.md`) itself — inline nothing — and returns \"two engineers would diverge here\" findings. The dispatch follows the contract in [`..\u002F_shared\u002Fagent-roster.md`](..\u002F_shared\u002Fagent-roster.md) (clean-isolated context, cited findings, `NO_AMBIGUITIES` if none). If `sdlc:devils-advocate` is unavailable at runtime, fall back to a `general-purpose` Agent with the prompt body in [`.\u002Freferences\u002Fambiguity-checks.md`](.\u002Freferences\u002Fambiguity-checks.md).\n\n4. **Merge + dedupe.** Union the self-sweep (step 2) with the subagent findings (step 3); collapse duplicates (same `§ref` + same class). Keep the highest-impact first (conflicting-requirement > under-specified-AC > unmeasured-NFR > undefined-term > missing-actor > scope-creep > vague-term > unstated-assumption). If the merged set is empty → report «PRD is unambiguous», stamp nothing, suggest `fix-term` \u002F `architecture-design`.\n\n5. **Resolve or defer, per finding.** For each finding, one `AskUserQuestion` phrased per [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md), offering: **Resolve now** (user picks\u002Fdictates the tightening) · **Defer to §8** (→ Open-Questions row, owner + due captured in a follow-up; missing either = stays unresolved, re-ask once) · **Not an ambiguity** (false positive — drop it, e.g. a term already in CONTEXT). Every finding ends Resolved or Deferred — none dangling. **Depth tunes the question volume, not the floor:** at `easy`, the skill resolves the unambiguous, low-stakes findings itself with sensible tightenings and lists them in a stated-assumptions ledger for a batch veto, asking only the behavioral\u002Fhigh-stakes forks; at `hard`, it asks every finding. The «zero dangling» rule holds at every level.\n\n6. **Write resolutions back.** Apply Resolve edits in the PRD's native section (tighten the §5 AC into business-observable form, replace a §6 adjective with a numeric target + measurement, fix a §1 term, add a missing §4 actor\u002FUS, cut a §3 scope-creep line). Append every Defer as a §8 checkbox row `- [ ] \u003Cquestion>? Default now: \u003CX>. — owner: \u003Cname\u002Frole>, due: \u003Cdate or stage>`. Keep a short edits-log (one line per finding: `class · §ref · resolved|deferred · before→after`).\n   - **Undefined-term findings reconcile the glossary in-flow (a hard rule, at every depth).** When an `undefined-term` finding is Resolved, immediately invoke `\u002Fsdlc:fix-term \u003Cslug>` for that term — compare it against `CONTEXT.md` and add\u002Fupdate the definition **now**, not as a deferred note. clarify-prd never resolves a term by inventing a meaning inline; the canonical definition lands in `CONTEXT.md` so `architecture-design` \u002F downstream stages read one source.\n\n7. **Stamp + commit.** Set `updated_at: \u003Ctoday>` in the frontmatter. Re-respect `write-prd`'s invariants when tightening (§5 AC stays free of HTTP\u002Fstatus\u002Ferror-code\u002FSQL tokens; §6 numbers carry a measurement; §4 roles only from the glossary). Propose `clarify-prd: \u003Cslug> — N resolved, M deferred`. Then **emit the stage-handoff block** per [`..\u002F_shared\u002Fhandoff.md`](..\u002F_shared\u002Fhandoff.md) — *What I did* + *Review* (tightened `PRD.md`) + *Run next* (`\u002Fclear`, then `\u002Fsdlc:fix-term \u003Cslug>`, or `\u002Fsdlc:architecture-design \u003Cslug>` to skip the glossary).\n\n## Definition of Done\n\n- Every ambiguity from the self-sweep + the subagent is **Resolved** (PRD tightened in its native section) or **Deferred** (a §8 row with both owner AND due) — zero dangling.\n- Resolve edits preserve `write-prd`'s contracts: §5 AC carry no HTTP\u002Fstatus\u002Ferror-code\u002FSQL tokens, §6 NFR rows carry a numeric target + measurement (no adjectives), §4 roles match the CONTEXT glossary.\n- The devil's-advocate subagent actually ran (clean context, Read the PRD itself) — not skipped, not paraphrased into the main thread.\n- `updated_at` reflects today; edits-log kept; commit proposed.\n\n## Anti-patterns\n\n- **Skipping the subagent** and sweeping only in-thread. The clean-context fork-finder is the core mechanic — the self-sweep is shaped by the conversation that wrote the PRD and misses its own blind spots.\n- **Resolving an ambiguity unilaterally** (no `AskUserQuestion`). clarify-prd proposes; the author decides — the same user-in-the-loop contract as every SDLC stage.\n- **Defer without owner+due** — a §8 row missing either is not a real defer; re-ask once, else it stays unresolved.\n- **Re-running the coherence critic here.** That F1–F6 cross-section drift check belongs to `write-prd` \u002F `architecture-design`; clarify-prd's subagent hunts *ambiguity* (build-divergence), a different target.\n- **Leaking implementation into a Resolve edit** — tightening a §5 AC by adding a status code \u002F endpoint \u002F SQL detail. Stay business-observable; the technical mapping lives in `api-forge` \u002F `generate-data-model`.\n- **Inventing answers to fill findings.** A genuinely open point is *deferred* with an owner, not guessed — better an honest §8 row than a fabricated AC.\n- **Authoring new scope.** clarify-prd sharpens what the PRD already says; a brand-new requirement goes back through `write-prd`, not in under cover of a \"clarification\".\n\n## References\n\n- [`.\u002Freferences\u002Fambiguity-checks.md`](.\u002Freferences\u002Fambiguity-checks.md) — the eight ambiguity classes (how to spot \u002F how to resolve each) + the clean-context devil's-advocate subagent prompt body.\n- [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md) — phrasing for the resolve\u002Fdefer\u002Ffalse-positive question.\n- [`..\u002F_shared\u002Finterview-depth.md`](..\u002F_shared\u002Finterview-depth.md) — the easy\u002Fmedium\u002Fhard dial set in step 1 (sweep aggressiveness + per-finding question volume).\n- [`..\u002F_shared\u002Fagent-roster.md`](..\u002F_shared\u002Fagent-roster.md) — model\u002Feffort policy + the shared agent dispatch contract.\n- [`..\u002F_shared\u002Fhandoff.md`](..\u002F_shared\u002Fhandoff.md) — the stage-handoff block format emitted at step 7.\n","# Skill: clarify-prd (SDLC stage 03 — PRD ambiguity sweep)\n\nAmbiguity sweep over a written `PRD.md`. It hunts the PRD for under-specified points — vague terms, unmeasured NFRs, AC missing error\u002Fauthz\u002Fedge behavior, unstated assumptions, conflicting requirements, undefined domain terms, missing actors, scope creep — then dispatches a **clean-context devil's-advocate subagent** that re-reads the PRD fresh and answers one question: *where would two engineers reasonably build different things from this?* Each ambiguity it surfaces is closed with the user: either **resolved** (the PRD is tightened in place) or **deferred** (a §8 Open-Questions row with owner + due). It exists so `fix-term` \u002F `architecture-design` never proceed on an ambiguous PRD.\n\nThis is a sweep, not a full authoring stage — it does **not** run the shared Socratic loop or the coherence critic. Its shared dependencies:\n→ [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md) (resolve\u002Fdefer question phrasing) · [`..\u002F_shared\u002Finterview-depth.md`](..\u002F_shared\u002Finterview-depth.md) (sweep aggressiveness + per-finding question volume) · [`..\u002F_shared\u002Fagent-roster.md`](..\u002F_shared\u002Fagent-roster.md) (dispatch discipline).\n\nDepth governs how aggressively the sweep hunts + the per-finding question volume. (Every surfaced ambiguity is still Resolved or Deferred at every level — that's a floor, not a dial.)\n\n## Owner\n\nPM + Tech Lead (the PRD's co-authors resolve their own ambiguities). PM owns vague-term \u002F scope \u002F missing-actor calls; Tech Lead owns unmeasured-NFR \u002F under-specified-AC \u002F conflicting-requirement calls.\n\n## Inputs\n\n- `\u003Cslug>` — same feature slug used by `write-prd`.\n- **Gate (hard-refuse if missing):** `docs\u002Ffeatures\u002F\u003Cslug>\u002FPRD.md`. Absent → STOP and point: «run `\u002Fsdlc:write-prd \u003Cslug>` first — clarify-prd sharpens an existing PRD, it does not write one».\n- (Optional) `docs\u002Ffeatures\u002F\u003Cslug>\u002FCONTEXT.md` — if present, its `## Glossary` is canonical; an \"undefined-term\" finding for a word already glossed there is a false positive (drop it).\n\n## Protocol\n\n1. **Gate + set interview depth.** `test -f docs\u002Ffeatures\u002F\u003Cslug>\u002FPRD.md` → missing = refuse with the pointer above. Read the PRD (and `CONTEXT.md` `## Glossary` if present, to suppress false \"undefined-term\" hits). **Then set the interview depth (the opening question):** read `interview_depth` from `.claude\u002Fsdlc.local.md` if present (else default medium), and — unless a `--depth=easy|medium|hard` arg was passed — ask ONE depth-selection `AskUserQuestion` phrased per [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md), with the saved\u002Fmedium value as the «(Recommended)» first option. The level tunes how adversarially the sweep + subagent hunt (easy: only build-divergence that changes behavior, with assumptions stated; medium: balanced; hard: adversarial, every fork surfaced) and the per-finding question volume → [`..\u002F_shared\u002Finterview-depth.md`](..\u002F_shared\u002Finterview-depth.md).\n\n2. **First-pass self-sweep.** Walk the PRD against the eight ambiguity classes in [`.\u002Freferences\u002Fambiguity-checks.md`](.\u002Freferences\u002Fambiguity-checks.md) (vague-term \u002F unmeasured-NFR \u002F under-specified-AC \u002F unstated-assumption \u002F conflicting-requirement \u002F undefined-term \u002F missing-actor \u002F scope-creep). Note candidate findings with a `§ref` each — do not edit yet.\n\n3. **Devil's-advocate subagent (the core mechanic).** Dispatch the `devils-advocate` agent — `subagent_type: \"sdlc:devils-advocate\"` (carries its own `model: opus` + `effort: high`; clean context — it never saw this conversation). Pass only the slug + the PRD path; it Reads `PRD.md` (and `CONTEXT.md`) itself — inline nothing — and returns \"two engineers would diverge here\" findings. The dispatch follows the contract in [`..\u002F_shared\u002Fagent-roster.md`](..\u002F_shared\u002Fagent-roster.md) (clean-isolated context, cited findings, `NO_AMBIGUITIES` if none). If `sdlc:devils-advocate` is unavailable at runtime, fall back to a `general-purpose` Agent with the prompt body in [`.\u002Freferences\u002Fambiguity-checks.md`](.\u002Freferences\u002Fambiguity-checks.md).\n\n4. **Merge + dedupe.** Union the self-sweep (step 2) with the subagent findings (step 3); collapse duplicates (same `§ref` + same class). Keep the highest-impact first (conflicting-requirement > under-specified-AC > unmeasured-NFR > undefined-term > missing-actor > scope-creep > vague-term > unstated-assumption). If the merged set is empty → report «PRD is unambiguous», stamp nothing, suggest `fix-term` \u002F `architecture-design`.\n\n5. **Resolve or defer, per finding.** For each finding, one `AskUserQuestion` phrased per [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md), offering: **Resolve now** (user picks\u002Fdictates the tightening) · **Defer to §8** (→ Open-Questions row, owner + due captured in a follow-up; missing either = stays unresolved, re-ask once) · **Not an ambiguity** (false positive — drop it, e.g. a term already in CONTEXT). Every finding ends Resolved or Deferred — none dangling. **Depth tunes the question volume, not the floor:** at `easy`, the skill resolves the unambiguous, low-stakes findings itself with sensible tightenings and lists them in a stated-assumptions ledger for a batch veto, asking only the behavioral\u002Fhigh-stakes forks; at `hard`, it asks every finding. The «zero dangling» rule holds at every level.\n\n6. **Write resolutions back.** Apply Resolve edits in the PRD's native section (tighten the §5 AC into business-observable form, replace a §6 adjective with a numeric target + measurement, fix a §1 term, add a missing §4 actor\u002FUS, cut a §3 scope-creep line). Append every Defer as a §8 checkbox row `- [ ] \u003Cquestion>? Default now: \u003CX>. — owner: \u003Cname\u002Frole>, due: \u003Cdate or stage>`. Keep a short edits-log (one line per finding: `class · §ref · resolved|deferred · before→after`).\n   - **Undefined-term findings reconcile the glossary in-flow (a hard rule, at every depth).** When an `undefined-term` finding is Resolved, immediately invoke `\u002Fsdlc:fix-term \u003Cslug>` for that term — compare it against `CONTEXT.md` and add\u002Fupdate the definition **now**, not as a deferred note. clarify-prd never resolves a term by inventing a meaning inline; the canonical definition lands in `CONTEXT.md` so `architecture-design` \u002F downstream stages read one source.\n\n7. **Stamp + commit.** Set `updated_at: \u003Ctoday>` in the frontmatter. Re-respect `write-prd`'s invariants when tightening (§5 AC stays free of HTTP\u002Fstatus\u002Ferror-code\u002FSQL tokens; §6 numbers carry a measurement; §4 roles only from the glossary). Propose `clarify-prd: \u003Cslug> — N resolved, M deferred`. Then **emit the stage-handoff block** per [`..\u002F_shared\u002Fhandoff.md`](..\u002F_shared\u002Fhandoff.md) — *What I did* + *Review* (tightened `PRD.md`) + *Run next* (`\u002Fclear`, then `\u002Fsdlc:fix-term \u003Cslug>`, or `\u002Fsdlc:architecture-design \u003Cslug>` to skip the glossary).\n\n## Definition of Done\n\n- Every ambiguity from the self-sweep + the subagent is **Resolved** (PRD tightened in its native section) or **Deferred** (a §8 row with both owner AND due) — zero dangling.\n- Resolve edits preserve `write-prd`'s contracts: §5 AC carry no HTTP\u002Fstatus\u002Ferror-code\u002FSQL tokens, §6 NFR rows carry a numeric target + measurement (no adjectives), §4 roles match the CONTEXT glossary.\n- The devil's-advocate subagent actually ran (clean context, Read the PRD itself) — not skipped, not paraphrased into the main thread.\n- `updated_at` reflects today; edits-log kept; commit proposed.\n\n## Anti-patterns\n\n- **Skipping the subagent** and sweeping only in-thread. The clean-context fork-finder is the core mechanic — the self-sweep is shaped by the conversation that wrote the PRD and misses its own blind spots.\n- **Resolving an ambiguity unilaterally** (no `AskUserQuestion`). clarify-prd proposes; the author decides — the same user-in-the-loop contract as every SDLC stage.\n- **Defer without owner+due** — a §8 row missing either is not a real defer; re-ask once, else it stays unresolved.\n- **Re-running the coherence critic here.** That F1–F6 cross-section drift check belongs to `write-prd` \u002F `architecture-design`; clarify-prd's subagent hunts *ambiguity* (build-divergence), a different target.\n- **Leaking implementation into a Resolve edit** — tightening a §5 AC by adding a status code \u002F endpoint \u002F SQL detail. Stay business-observable; the technical mapping lives in `api-forge` \u002F `generate-data-model`.\n- **Inventing answers to fill findings.** A genuinely open point is *deferred* with an owner, not guessed — better an honest §8 row than a fabricated AC.\n- **Authoring new scope.** clarify-prd sharpens what the PRD already says; a brand-new requirement goes back through `write-prd`, not in under cover of a \"clarification\".\n\n## References\n\n- [`.\u002Freferences\u002Fambiguity-checks.md`](.\u002Freferences\u002Fambiguity-checks.md) — the eight ambiguity classes (how to spot \u002F how to resolve each) + the clean-context devil's-advocate subagent prompt body.\n- [`..\u002F_shared\u002Fask-style.md`](..\u002F_shared\u002Fask-style.md) — phrasing for the resolve\u002Fdefer\u002Ffalse-positive question.\n- [`..\u002F_shared\u002Finterview-depth.md`](..\u002F_shared\u002Finterview-depth.md) — the easy\u002Fmedium\u002Fhard dial set in step 1 (sweep aggressiveness + per-finding question volume).\n- [`..\u002F_shared\u002Fagent-roster.md`](..\u002F_shared\u002Fagent-roster.md) — model\u002Feffort policy + the shared agent dispatch contract.\n- [`..\u002F_shared\u002Fhandoff.md`](..\u002F_shared\u002Fhandoff.md) — the stage-handoff block format emitted at step 7.\n",{"name":5,"model":24,"stage":25,"agents":26,"effort":28,"triggers":29,"description":7},"opus","03",[27],"devils-advocate","high",[30,31,32,33,34],"\u002Fsdlc:clarify-prd","clarify","find ambiguities in","is the PRD ready","sharpen the PRD","main","2dc9bdfa5d508a04d414fb0991617e6722ec635a","modules\u002F6-sdlc\u002Fsdlc\u002Fplugin\u002Fskills\u002Fclarify-prd",13,"https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public","https:\u002F\u002Ft.me\u002Fgenkovich","https:\u002F\u002Fgithub.com\u002Fgenkovich","https:\u002F\u002Ft.me\u002Fbeerphp","Beer::PHP 🍺 telegram-channel author \r\n","2026-11-12 03:03:13",{"source":8,"skillName":5,"commands":46,"prompt":56,"zipAvailable":57},[47,50,53],{"tool":48,"command":49},"npx","npx skills add genkovich\u002Fagentic-engineering-course-public --skill clarify-prd",{"tool":51,"command":52},"bunx","bunx skills add genkovich\u002Fagentic-engineering-course-public --skill clarify-prd",{"tool":54,"command":55},"pnpm","pnpm dlx skills add genkovich\u002Fagentic-engineering-course-public --skill clarify-prd","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F6-sdlc\u002Fsdlc\u002Fplugin\u002Fskills\u002Fclarify-prd\u002FSKILL.md",true,[59],{"path":60,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":63,"sourceHtmlUrl":64},"references\u002Fambiguity-checks.md","REFERENCE","参考文档",9467,"https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F6-sdlc\u002Fsdlc\u002Fplugin\u002Fskills\u002Fclarify-prd\u002Freferences\u002Fambiguity-checks.md",[66,74,81,88,95,102],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":73},"01M72YGBW5JKECT8WSRHK1YBCJ","release","Run the whole local release pipeline in order, stopping at each human gate. Triggers on \"run the release\", \"prepare a release\", \"\u002Frelease\", \"зроби реліз\", \"підготуй реліз\". Drives the four stages back to back — bump-version → curate-changelog → release-notes → check-docs-drift — pausing after each so the human reviews the diff before the next stage. Prepares everything in the working tree; never tags, commits, or pushes. The tag push (which fires CI) stays the human's action.","https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F9-collaboration\u002F9.7-release-docs\u002F.claude\u002Fskills\u002Frelease\u002FSKILL.md",[],[],"2026-11-11 15:11:14",{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":73},"01M72YG87VQ89PRCYQSMNAP5RW","release-notes","Draft partner-facing release notes from the same commits as the changelog. Triggers on \"draft release notes\", \"release notes for the next version\", \"\u002Frelease-notes\", \"нотатки релізу\", \"release notes партнеру\". Same input as curate-changelog, different audience and tone: a short, benefit-first narrative for a user or partner — \"what's new for you\" — not a categorized engineer's list. Keeps only changes a user actually notices; drops internal refactors and tooling. Prints the notes to the chat to seed a GitHub Release description or a partner announcement. Writes no file.","https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F9-collaboration\u002F9.7-release-docs\u002F.claude\u002Fskills\u002Frelease-notes\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":73},"01M72YG4BXJH9R6QQZD3G8E358","curate-changelog","Curate the [Unreleased] section of docs\u002FCHANGELOG.md from git history. Triggers on \"curate the changelog\", \"update the changelog\", \"\u002Fcurate-changelog\", \"онови changelog\", \"курування журналу\". Reads git log since the last tag, FILTERS out what a reader never feels (test\u002Fci\u002Fchore\u002Fmerge), GROUPS commits about one feature into one bullet, and REWRITES technical subjects into plain language under the six Keep a Changelog categories (Added\u002FChanged\u002FDeprecated\u002F Removed\u002FFixed\u002FSecurity). The result must have FEWER lines than the raw log — that is the proof of curation, not a dump. Leaves the edit in the working tree; never commits or pushes.","https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F9-collaboration\u002F9.7-release-docs\u002F.claude\u002Fskills\u002Fcurate-changelog\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":73},"01M72YFZDKNXQ7WF9ARC5KQ4F5","codify-rule","Turn a repeated mistake into a durable, path-scoped rule in .claude\u002Frules\u002F. Triggers on \"codify this rule\", \"make a rule so this doesn't happen again\", \"\u002Fcodify-rule\", \"закодифікуй правило\", \"петля помилка-правило\". Use only when the same mistake showed up TWICE (twice = a pattern, once = noise). Names the wrong shape and the right shape with one short example of each, scopes the rule to the paths that matter, and leaves the edit in the working tree so the next agent on this repo inherits the fix. Edits only .claude\u002Frules\u002F; never commits or pushes.","https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F9-collaboration\u002F9.7-release-docs\u002F.claude\u002Fskills\u002Fcodify-rule\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":73},"01M72YFVKC19NCQDD30R1C282Q","check-docs-drift","Find and fix where docs\u002Fapi.md fell behind the code. Triggers on \"check for doc drift\", \"is the API doc up to date\", \"\u002Fcheck-docs-drift\", \"перевір дрейф документації\", \"docs drift\". Runs the deterministic scripts\u002Fcheck-docs-drift.py (parses the public NoteBook methods with ast, compares to what docs\u002Fapi.md documents — NO LLM finds the gap), reports the drift, and PROPOSES the doc edit that closes it. The gap is found by parsing; the agent writes the missing documentation in the doc's existing style. Edits only docs\u002F; never commits or pushes.","https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F9-collaboration\u002F9.7-release-docs\u002F.claude\u002Fskills\u002Fcheck-docs-drift\u002FSKILL.md",[],[],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":73},"01M72YFQSJ21QX4SQADQ1SQDZR","bump-version","Decide and apply the next semver for a release. Triggers on \"bump the version\", \"what version is next\", \"\u002Fbump-version\", \"простав версію\", \"яка наступна версія\". Reads the next number from the deterministic scripts\u002Fnext-version.sh (NEVER guesses it), explains in plain language WHICH part moved and why (e.g. \"MINOR, because a feat landed: filter notes by tag\"), edits the version line in pyproject.toml, and proposes the matching `git tag vX.Y.Z`. The number is decided by a rule; the agent explains and applies — the human confirms the tag. Never pushes a tag itself.","https:\u002F\u002Fgithub.com\u002Fgenkovich\u002Fagentic-engineering-course-public\u002Fblob\u002Fmain\u002Fmodules\u002F9-collaboration\u002F9.7-release-docs\u002F.claude\u002Fskills\u002Fbump-version\u002FSKILL.md",[],[]]