[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79QH32MD4W355P2MNBDHE89":3,"skill-same:01M79QH32MD4W355P2MNBDHE89":61},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":25,"commitSha":26,"skillDirectory":27,"forksCount":13,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M79QH32MD4W355P2MNBDHE89","plugin-conception",null,"Turns a rough idea, imported draft, or recent-session evidence of friction, failure, or opportunity into a decision-ready classification (Create, Enhance, Repair, Consolidate, Reposition, Retain, Reject\u002FDefer) before plugin-ideation's interview or plugin-planning\u002FFix begins. Runs a shared normalize\u002Fclassify\u002Fshallow-overlap-check core from either starting point, then produces a Conception Brief whose depth matches the classification — light for Create (hands off to plugin-ideation unchanged), full for every other outcome (marketplace integration, implementation plan, baseline contract). Use when the user asks \"should this be a new plugin or an enhancement\", \"classify this idea\", \"is this new or should it extend X\", \"turn these session findings into a concept\", or has a rough idea or session evidence that needs qualifying before ideation or planning\u002FFix runs. Not a replacement for plugin-ideation's interview or plugin-planning's component inventory — this is the qualifier that precedes both.","AndreHahm\u002Fandres-cc-marketplace","AndreHahm","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F80748019?v=4","01M79N2VZJT9MAJ7K695C9VXQ4","01M79N2W06PFX8BSCQ592M8K8K",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-devkit\u002Fskills\u002Fplugin-conception\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: plugin-conception\ndescription: >-\n  Turns a rough idea, imported draft, or recent-session evidence of friction, failure, or\n  opportunity into a decision-ready classification (Create, Enhance, Repair, Consolidate,\n  Reposition, Retain, Reject\u002FDefer) before plugin-ideation's interview or plugin-planning\u002FFix\n  begins. Runs a shared normalize\u002Fclassify\u002Fshallow-overlap-check core from either starting point,\n  then produces a Conception Brief whose depth matches the classification — light for Create\n  (hands off to plugin-ideation unchanged), full for every other outcome (marketplace\n  integration, implementation plan, baseline contract). Use when the user asks \"should this be a\n  new plugin or an enhancement\", \"classify this idea\", \"is this new or should it extend X\", \"turn\n  these session findings into a concept\", or has a rough idea or session evidence that needs\n  qualifying before ideation or planning\u002FFix runs. Not a replacement for plugin-ideation's\n  interview or plugin-planning's component inventory — this is the qualifier that precedes both.\nargument-hint: \"[rough idea, problem statement, or session-evidence source]\"\nallowed-tools: Read Glob Grep AskUserQuestion Write Bash(date:*) Skill\n---\n\n# Plugin Conception\n\nTurns either a rough idea or evidence of an unmet need into a decision-ready classification before\n`plugin-ideation`'s interview, `plugin-planning`'s component inventory, or `plugin-lifecycle-maintenance`'s\nFix phase begins. This is a **pre-ideation, pre-planning qualifier**, not a replacement for the deeper work\neither downstream skill already does — it establishes that a proposed change is useful, distinct,\nappropriately placed, bounded, and worth the next phase's cost, then hands off.\n\n**Conception is a classification pass, not an interview.** Unlike `plugin-ideation`'s multi-round\nback-and-forth, this skill interviews the requester only where a missing answer would change the\nclassification or the shallow overlap read — never to `plugin-ideation`'s depth, and never re-asking\nquestions `plugin-ideation` will ask again immediately after a Create handoff.\n\n## Quick Start\n\n1. **Route** — from-scratch idea, or recent-session evidence? (Entry Routes A\u002FB)\n2. **Normalize and classify** — build a problem frame, label assumptions, then classify: Create \u002F Enhance \u002F Repair \u002F Consolidate \u002F Reposition \u002F Retain \u002F Reject-Defer (Steps 1-2)\n3. **Check overlap, shallow only** — repository-metadata depth; confirms or revises the classification (Step 3)\n4. **Define the concept, non-Create only** — full depth, implementation plan, workflow if procedural (Steps 4-6; Create skips straight to Step 7)\n5. **Decide and hand off** — approve \u002F revise \u002F merge \u002F defer \u002F reject; write the Conception Brief (Step 7)\n\n## When to Use\n\n- Before `plugin-ideation`'s interview, to confirm a rough idea is actually worth that interview (or is\n  really an Enhance\u002FConsolidate in disguise)\n- Before `plugin-planning`\u002F`plugin-lifecycle-maintenance`'s Fix phase, to turn session evidence, a finding,\n  or an enhancement request into a bounded, evidence-backed concept with a baseline contract\n- Invoked automatically as Phase 1 (Conceive) of `plugin-lifecycle-upstream`, or as a step inside\n  `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin` workflows, immediately after their\n  own human finding-selection gate\n\n## When NOT to Use\n\n- The concept is already a Create classification with a clear problem statement and nothing left to\n  qualify — go straight to `plugin-ideation`; running this skill first only to immediately hand off is\n  needless ceremony when the classification was never in doubt\n- A narrow repair with an already-known fix and an already-accepted finding — see \"Bypass for narrow\n  repairs\" under Step 2; hand off directly to Fix instead\n- Deep per-component overlap search, activation-trigger overlap analysis, or naming-candidate validation —\n  that's `plugin-ideation` Step 3-4's job (Create) or the type-matched reviewer agents' job (everything\n  else), never this skill's\n- Scaffolding components, editing production plugins, or running QA fixes — this skill produces a brief,\n  never files\n- Already committed to running the full guided pipeline (Conceive→Ideate→Plan→Design→Build), not just the\n  classification step — use `plugin-lifecycle-upstream` instead; it dispatches here automatically as\n  Phase 1, so invoking this skill directly is for when the classification decision itself is the goal\n- Already committed to acting on retro\u002Fcomparison findings against an already-shipped plugin, with\n  apply\u002Ftest\u002Fcommit follow-through — use `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F\n  `enhance-a-plugin` workflows instead; they dispatch here automatically after the human finding-selection\n  pick, so invoking this skill directly is for classifying a rough idea or evidence source before any\n  pipeline commitment exists\n\n## Entry Route A: Start From Scratch\n\nUse when the starting point is a rough idea, problem statement, desired capability, or proposed\nplugin\u002Fskill. If `$ARGUMENTS` is given, treat it as the rough idea or problem statement directly — the\nsame way `plugin-ideation`'s own Step 2 treats a detailed `$ARGUMENTS` — and reflect it back rather than\nre-asking for it from scratch. Minimum intake, deliberately light since a Create outcome hands off to\n`plugin-ideation`'s own deeper interview rather than duplicating it here:\n\n- the problem or opportunity;\n- the concrete user and usage situation, at whatever resolution the requester already has it;\n- the intended outcome;\n- known constraints and explicit non-goals, as far as they're already known;\n- any existing component the requester believes is adjacent.\n\nInterview the requester only where a missing answer changes classification or the shallow overlap read.\nDo not fabricate evidence that does not exist, and do not require a retrospective the requester never\nasked for.\n\n## Entry Route B: Start From Recent-Session Evidence\n\nUse when repeated friction, failures, gaps, or opportunities have been observed in one or more recent\nsessions — session-analysis reports, build handoff reports, validation\u002Fgrading\u002Fcomparison\u002Fsecurity\u002F\ncompleteness\u002Fconsistency findings, user corrections, or durable planning artifacts. If `$ARGUMENTS` names\none of these sources directly (a report path, or one entry from a findings bundle or approved suggestion\nlist an upstream workflow's own human-decision gate already produced), treat it as the evidence source for\nStep 1 below rather than re-deriving it. **This skill classifies and briefs exactly one candidate per\ninvocation** — when a caller's own gate approved several candidates at once (e.g.\n`plugin-lifecycle-maintenance`'s multiSelect finding-selection pick), it invokes this skill once per\ncandidate, never once with the whole list; this skill's own Step 1 never fans a single invocation out into\nmultiple classifications or multiple briefs. Session evidence is a **seed, not the scope** of the concept.\nSee `references\u002Fevidence-routing.md` for the full 6-step\nevidence-handling procedure (identify source, recheck currency, separate symptoms from underlying need,\nmerge duplicates, discard stale\u002Fnon-actionable, obtain explicit approval before promoting to a planned\nchange) and the evidence-source list. If evidence is insufficient, fall back to Entry Route A's focused\ninterview — never require a full retrospective the desired concept doesn't actually depend on.\n\nMost session evidence classifies as Enhance, Repair, or Consolidate rather than Create — those\nclassifications get the full Conception Brief (Steps 4-5 below), since `plugin-lifecycle-maintenance`'s\nexisting workflows have no equivalent deep-definition step of their own.\n\n## Step 1: Normalize Intent and Evidence\n\nBuild a concise problem frame from whichever entry route ran:\n\n- problem or opportunity;\n- target user and concrete scenario;\n- desired outcome;\n- supporting evidence or assumptions;\n- constraints and non-goals;\n- success signals.\n\nLabel every assumption explicitly. Every piece of evidence must identify its source. Never rely on an\nunverified claim that a capability is absent or broken — check the current repository state instead of\ntrusting a stale report.\n\n## Step 2: Classify the Proposed Change\n\n| Classification | Meaning | Likely destination |\n|---|---|---|\n| Create | No suitable existing capability; add a new plugin or component | `plugin-ideation`, then the rest of the upstream pipeline |\n| Enhance | Preserve an existing component and add or improve behavior | `plugin-planning` (if new components are needed) or directly to Fix |\n| Repair | Restore behavior that contradicts the current contract | Directly to Fix — usually skips the full brief (see below) |\n| Consolidate | Multiple components overlap and should be merged or have clearer boundaries | `plugin-planning` plus consistency\u002Factivation review |\n| Reposition | Capability exists but belongs in a different plugin or functional group | `plugin-planning` and marketplace restructuring |\n| Retain | Current behavior is adequate; no implementation is justified | Stop with rationale |\n| Reject\u002FDefer | Benefit, evidence, feasibility, or priority is insufficient | Stop or record a revisit condition |\n\nThe classification may change after Step 3's overlap check — an apparent Create may turn out to be Enhance\nor Consolidate, and, just as often, an apparent Enhance may turn out to be a genuine Create once the\nshallow check shows nothing adjacent actually exists.\n\n**Bypass for narrow repairs:** a Repair classification with an already-known, narrowly-scoped fix and an\nalready-accepted finding does not need the full brief (Steps 4-5) — record the classification and\nevidence, then hand off directly to Fix. This keeps the skill proportional rather than adding ceremony to\na one-line fix.\n\n## Step 3: Check Overlap and Marketplace Fit — Shallow, Classification-Level Only\n\nThis check is deliberately shallow. Its only job is to confirm or revise Step 2's classification — never\nthe exhaustive per-component search `plugin-ideation` runs immediately after a Create handoff, and never\nthe deep evidence-gathering the finding-source tools already ran before this skill started.\n\nInspect, at repository-metadata depth only:\n\n- plugin manifests and plugin purposes;\n- skill, agent, command, hook, and rule descriptions;\n- existing lifecycle ownership;\n- related draft concepts and generated reports.\n\n| Overlap | Meaning |\n|---|---|\n| None | Step 2's classification likely holds; proceed |\n| Partial | Classification may need to shift toward Enhance\u002FConsolidate; reconsider before proceeding |\n| Full | Classification shifts to Retain or Enhance; a Create outcome is very unlikely to survive this finding |\n\n**Do not, here:** activation-trigger overlap analysis, cross-component reference tracing, or\nnaming-candidate validation — escalating past metadata-level inspection is the single most common way this\nskill ends up duplicating `plugin-ideation`. Marketplace placement follows the repository convention: a\nplugin uses `\u003Cdomain>-kit` or `\u003Cdomain>-devkit`, exactly one hyphen immediately before the suffix. For\nCreate outcomes, this skill never proposes a name — that's `plugin-ideation` Step 4's job. For every other\noutcome, record the existing plugin's placement, not a new name.\n\n**If the classification settled here is Retain, Reject, or Defer:** stop now, per Stop Conditions below —\ndo not proceed to Steps 4-7, and do not write a Conception Brief file. State the classification and\nrationale directly in the response instead. This is the only path that produces \"no brief written\" — see\nthe Testing & Validation section's Retain scenario.\n\n## Step 4: Define the Concept — Full Depth Only for Non-Create\n\n**Create:** skip this step entirely. Step 1's problem frame plus the classification and Step 3's shallow\noverlap result is the whole Create handoff payload.\n\n**Enhance \u002F Repair (full-brief) \u002F Consolidate \u002F Reposition:** the accepted concept must specify a\none-sentence value proposition, target users and triggering situations, in-scope capabilities, explicit\nnon-goals, whether the target is a new component or a change to an existing one (never a new plugin — that\nwould be a Create by definition), the proposed marketplace owner and neighboring components,\ndependencies\u002Fconsumers\u002Flifecycle handoffs, compatibility requirements for existing behavior, measurable\nacceptance criteria, principal risks and mitigations, and a baseline contract (behavior that must remain\nunchanged, observed deficiency and evidence, proposed delta, affected callers\u002Fdocs\u002Ftests\u002Factivation\ntriggers, rollback or rejection conditions).\n\n## Step 5: Shape the Implementation Plan\n\n**Create:** skip this step — `plugin-planning` (reached after `plugin-ideation` completes) owns\ncomponent-type\u002Fcount\u002Fdepth decisions.\n\n**Every other non-bypassed classification:** define target files\u002Fcomponent directories, required new or\nmodified components, dependency\u002Fintegration changes, tests or evaluations that demonstrate the concept,\ndocumentation changes, validation\u002Freview\u002Fcommit boundaries, and a recommended execution order. Detailed\nprompts, scripts, schemas, and line-level edits still belong in the subsequent planning\u002Fdesign phase — not\nhere.\n\n## Step 6: Describe the Workflow When the Concept Is Procedural\n\nInclude a workflow only when the proposed capability coordinates three or more dependent states, gates, or\nhandoffs — entry conditions, ordered phases, artifacts consumed\u002Fproduced, human approval gates, stop\u002F\nretry\u002Fresume behavior, downstream owner per handoff, completion criteria. A single focused skill does not\nneed an invented workflow merely to fill this section — mark it not applicable with a one-sentence reason\ninstead.\n\n## Step 7: Decide and Hand Off\n\n**Entry:** the classification is Create, Enhance, Repair (full-brief path), Consolidate, or Reposition — a\nRetain\u002FReject\u002FDefer classification already stopped before reaching this step (see the gate at the end of\nStep 3).\n\nPresent the completed brief via `AskUserQuestion`: approve and proceed \u002F revise \u002F merge with an existing\nconcept \u002F defer with a revisit condition \u002F reject or retain the current state instead. This last pair of\noptions is the user overriding an already-written brief post-hoc (e.g. approving the classification but\ndeciding on reflection the change isn't worth doing) — a different case from the classification-level\nRetain\u002FReject\u002FDefer stop above, which never reaches this step at all. Only an approved concept proceeds —\nthis skill never auto-applies findings or silently turns evidence into a feature.\n\n1. Get a timestamp: `date -u +%Y-%m-%dT%H-%M-%SZ`\n2. Write to `.claude\u002Foutput\u002Fplugin-conception\u002F\u003Cconcept-slug>-\u003Ctimestamp>.md` per\n   `references\u002Fconception-brief-template.md` — for Create, the light variant its own blockquote\n   instructs (the section list lives there, not restated here, to avoid drift); for Enhance\u002FRepair\n   (full-brief)\u002FConsolidate\u002FReposition, all 12 sections in full\n3. Present the artifact link first, then the summary:\n\n```\n📄 Conception Brief written: `.claude\u002Foutput\u002Fplugin-conception\u002F\u003Cconcept-slug>-\u003Ctimestamp>.md`\n```\n\nHand-off target follows the classification:\n\n- **Create →** `plugin-ideation`, with the light brief as its input.\n- **Enhance \u002F Consolidate \u002F Reposition →** `plugin-planning` if new or restructured components are\n  implied, otherwise directly to `plugin-lifecycle-downstream`'s Phase 8 (Consolidated Fix).\n- **Repair (full-brief path) →** directly to `plugin-lifecycle-downstream`'s Phase 8 (Consolidated Fix).\n- **Retain \u002F Reject \u002F Defer →** stop; no downstream hand-off.\n\n**Phase 8 needs a findings bundle, not a brief — this skill never invokes it directly.**\n`plugin-lifecycle-downstream`'s Phase 8 external-entry contract accepts only a schema-validated findings\nbundle (`plugin-rulebook\u002Freferences\u002Fevidence-schema.md`'s Finding\u002FReport-Revision shape, with a\n`baseline_commit` and scope manifest) and refuses anything else — a Conception Brief path does not\nsatisfy it, and this skill has no `Bash(git log:*)` grant to produce `baseline_commit` even if it tried.\nWhere Phase 8 is named as the destination above, the reformatting-and-invoking step belongs to whichever\norchestrator called this skill (e.g. `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin`\nStep 4 already does this conversion) — never to this skill itself.\n\n**Standalone invocation:** for a `plugin-ideation`\u002F`plugin-planning` hand-off, ask via `AskUserQuestion`\nbefore invoking the target — never invoke it silently. For a Phase 8 destination, there is no orchestrator\npresent to do the conversion: state the classification and that a findings-bundle conversion is needed\nbefore Phase 8 can be reached, and stop — never attempt to invoke `plugin-lifecycle-downstream` directly.\n\n**Nested invocation:** when this skill runs as a step inside another orchestrator — `plugin-lifecycle-upstream`'s\nPhase 1 (Conceive), or `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin` Conceive step —\nthe brief-content approval above (approve\u002Frevise\u002Fmerge\u002Fdefer\u002Freject) still runs, but the hand-off\n**invocation** does not: this step ends at \"write the brief and report its classification.\" The calling\norchestrator owns deciding where each classified candidate goes next, since it may be routing several\ncandidates from one run — that decision may be a single consolidated `AskUserQuestion` (as\n`plugin-lifecycle-upstream`'s Gate 1 does) or a deterministic routing rule stated plainly to the user with\nno further ask (as `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin` Step 4 does,\nrouting by classification alone). Either way, never ask twice for the same decision — see those two\ncallers for how each owns this instead.\n\n## Resume Behavior\n\n- A draft brief resumes at the first incomplete section (light briefs have fewer sections to resume\n  through).\n- An approved brief resumes at handoff after the user confirms it is still current.\n- Evidence is rechecked when the underlying target has changed since approval.\n- A legacy Concept Card may still be read directly by `plugin-ideation` without ever passing through this\n  skill — conception is not a mandatory gate in front of an already-existing Concept Card.\n\n## Stop Conditions\n\nStop without implementation when: full overlap makes a new component unnecessary; current behavior\nsatisfies the verified need; evidence is stale or already resolved; the concept lacks a concrete user\noutcome; marketplace ownership cannot be decided without a user choice; risks or cost exceed the accepted\nbenefit; or the user defers or rejects the concept. A clean stop is a valid result and must include its\nrationale — never silently drop a concept without stating why.\n\n## Testing & Validation\n\n**Eval evidence:** `evals\u002Fplugin-conception\u002Fevals.json` — 8 scenarios (Quick Workflow, `workspace\u002Fiteration-1`), 7\u002F8 eval-covered (scenario 1's eval is invalidated by a scripted-premise flaw unrelated to this skill's behavior — see its own `grading.json`); the remaining scenarios below are design-review-verified.\n\n1. **From-scratch Create** — a rough idea with no session evidence; confirm the light brief (per\n   `references\u002Fconception-brief-template.md`'s own light-variant section list) is produced and handed to\n   `plugin-ideation`, never a padded-out full brief\n2. **Session-evidence Enhance** — evidence pointing at friction with an existing component; confirm the\n   full brief is produced (all 12 sections) with a populated baseline contract\n3. **Full overlap** — an idea that duplicates an existing component; confirm the classification shifts to\n   Retain\u002FEnhance and the skill stops with rationale rather than proceeding toward Create\n4. **Stale evidence** — session evidence describing a since-fixed gap; confirm Step 1's current-state\n   reverification catches this and the evidence is discarded, not promoted\n5. **Narrow-repair bypass** — a Repair classification with an already-known fix and an accepted finding;\n   confirm Steps 4-5 are skipped and the skill hands off directly to Fix\n6. **Consolidate** — multiple overlapping components found during Step 3; confirm the full brief is\n   produced and `plugin-planning` plus consistency\u002Factivation review are named as the destination\n7. **Retain\u002Fno-work outcome** — evidence that current behavior is already adequate; confirm a clean stop\n   with stated rationale, no brief written\n8. **Create brief stays light** — construct a Create classification with rich session evidence available;\n   confirm the written brief still marks Scope and Boundaries, Existing-Component Baseline, Implementation\n   Plan, and the \"eventual implementation success\" half of Acceptance Criteria as not applicable, rather\n   than restating what `plugin-ideation`\u002F`plugin-planning` will produce\n\n**Quality gates:**\n- [ ] Step 3's overlap check never exceeds repository-metadata depth — no activation-trigger analysis,\n      cross-component reference tracing, or naming-candidate validation\n- [ ] A Create classification's brief always marks Scope and Boundaries, Existing-Component Baseline,\n      Implementation Plan, and the \"eventual implementation success\" half of Acceptance Criteria not\n      applicable — never a padded-out 12-section document\n- [ ] The narrow-repair bypass (Step 2) is only ever taken when the fix is already known and the finding\n      already accepted — never for an undefined or contested repair\n- [ ] Entry Route B evidence is always rechecked against current repository state before promotion — never\n      copied forward from a source artifact unverified\n- [ ] Step 7's decision is always made via `AskUserQuestion` — this skill never auto-applies a finding or\n      silently promotes evidence into a planned change\n- [ ] **Standalone invocation:** the hand-off target is always invoked via `AskUserQuestion` confirmation\n      first — never invoked silently. **Nested invocation** (inside `plugin-lifecycle-upstream`'s Phase 1\n      or `plugin-lifecycle-maintenance`'s Conceive step): no hand-off is invoked from here at all — only\n      the brief-content approval fires; the calling orchestrator owns the hand-off routing decision (a\n      consolidated ask, or a plainly-stated deterministic rule)\n- [ ] The Conception Brief path is always under `.claude\u002Foutput\u002Fplugin-conception\u002F`\n- [ ] A clean stop (Retain\u002FReject\u002FDefer\u002Ffull overlap\u002Fstale evidence) always states its rationale — never a\n      silent drop\n- [ ] A classification-level Retain\u002FReject\u002FDefer outcome always stops at the end of Step 3, before Steps\n      4-7 run — no Conception Brief file is ever written for it\n\n## Reference Guide\n\n| Resource | Type | Purpose |\n|---|---|---|\n| `scripts\u002Fsmoke_test.py` | In this skill | This skill's own persisted smoke test (frontmatter validity, referenced-file existence, Bash-scope grant consistency) — re-run after any SKILL.md edit |\n| `evals\u002Fplugin-conception\u002F` | Repo root | Persisted `skill-tester` Quick Workflow eval suite (8 scenarios, 7\u002F8 covered) |\n| `references\u002Fconception-brief-template.md` | In this skill | Light and full Conception Brief templates used in Step 7 |\n| `references\u002Fevidence-routing.md` | In this skill | Entry Route B's evidence sources and the 6-step evidence-handling procedure |\n| `plugin-ideation` skill | Sibling skill | Create-classification hand-off target — owns the interview, deep overlap search, and naming |\n| `plugin-planning` skill | Sibling skill | Hand-off target for Enhance\u002FConsolidate\u002FReposition outcomes implying new or restructured components |\n| `plugin-lifecycle-downstream` skill | Sibling skill | Owns Phase 8 (Consolidated Fix) — hand-off target for Repair and no-new-component Enhance outcomes |\n| `plugin-lifecycle-upstream` skill | Sibling skill | Runs this skill as Phase 1 (Conceive) |\n| `plugin-lifecycle-maintenance` skill | Sibling skill | Runs this skill as a step inside `improve-a-plugin`\u002F`enhance-a-plugin`, after the human finding-selection gate |\n","# Plugin Conception\n\nTurns either a rough idea or evidence of an unmet need into a decision-ready classification before\n`plugin-ideation`'s interview, `plugin-planning`'s component inventory, or `plugin-lifecycle-maintenance`'s\nFix phase begins. This is a **pre-ideation, pre-planning qualifier**, not a replacement for the deeper work\neither downstream skill already does — it establishes that a proposed change is useful, distinct,\nappropriately placed, bounded, and worth the next phase's cost, then hands off.\n\n**Conception is a classification pass, not an interview.** Unlike `plugin-ideation`'s multi-round\nback-and-forth, this skill interviews the requester only where a missing answer would change the\nclassification or the shallow overlap read — never to `plugin-ideation`'s depth, and never re-asking\nquestions `plugin-ideation` will ask again immediately after a Create handoff.\n\n## Quick Start\n\n1. **Route** — from-scratch idea, or recent-session evidence? (Entry Routes A\u002FB)\n2. **Normalize and classify** — build a problem frame, label assumptions, then classify: Create \u002F Enhance \u002F Repair \u002F Consolidate \u002F Reposition \u002F Retain \u002F Reject-Defer (Steps 1-2)\n3. **Check overlap, shallow only** — repository-metadata depth; confirms or revises the classification (Step 3)\n4. **Define the concept, non-Create only** — full depth, implementation plan, workflow if procedural (Steps 4-6; Create skips straight to Step 7)\n5. **Decide and hand off** — approve \u002F revise \u002F merge \u002F defer \u002F reject; write the Conception Brief (Step 7)\n\n## When to Use\n\n- Before `plugin-ideation`'s interview, to confirm a rough idea is actually worth that interview (or is\n  really an Enhance\u002FConsolidate in disguise)\n- Before `plugin-planning`\u002F`plugin-lifecycle-maintenance`'s Fix phase, to turn session evidence, a finding,\n  or an enhancement request into a bounded, evidence-backed concept with a baseline contract\n- Invoked automatically as Phase 1 (Conceive) of `plugin-lifecycle-upstream`, or as a step inside\n  `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin` workflows, immediately after their\n  own human finding-selection gate\n\n## When NOT to Use\n\n- The concept is already a Create classification with a clear problem statement and nothing left to\n  qualify — go straight to `plugin-ideation`; running this skill first only to immediately hand off is\n  needless ceremony when the classification was never in doubt\n- A narrow repair with an already-known fix and an already-accepted finding — see \"Bypass for narrow\n  repairs\" under Step 2; hand off directly to Fix instead\n- Deep per-component overlap search, activation-trigger overlap analysis, or naming-candidate validation —\n  that's `plugin-ideation` Step 3-4's job (Create) or the type-matched reviewer agents' job (everything\n  else), never this skill's\n- Scaffolding components, editing production plugins, or running QA fixes — this skill produces a brief,\n  never files\n- Already committed to running the full guided pipeline (Conceive→Ideate→Plan→Design→Build), not just the\n  classification step — use `plugin-lifecycle-upstream` instead; it dispatches here automatically as\n  Phase 1, so invoking this skill directly is for when the classification decision itself is the goal\n- Already committed to acting on retro\u002Fcomparison findings against an already-shipped plugin, with\n  apply\u002Ftest\u002Fcommit follow-through — use `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F\n  `enhance-a-plugin` workflows instead; they dispatch here automatically after the human finding-selection\n  pick, so invoking this skill directly is for classifying a rough idea or evidence source before any\n  pipeline commitment exists\n\n## Entry Route A: Start From Scratch\n\nUse when the starting point is a rough idea, problem statement, desired capability, or proposed\nplugin\u002Fskill. If `$ARGUMENTS` is given, treat it as the rough idea or problem statement directly — the\nsame way `plugin-ideation`'s own Step 2 treats a detailed `$ARGUMENTS` — and reflect it back rather than\nre-asking for it from scratch. Minimum intake, deliberately light since a Create outcome hands off to\n`plugin-ideation`'s own deeper interview rather than duplicating it here:\n\n- the problem or opportunity;\n- the concrete user and usage situation, at whatever resolution the requester already has it;\n- the intended outcome;\n- known constraints and explicit non-goals, as far as they're already known;\n- any existing component the requester believes is adjacent.\n\nInterview the requester only where a missing answer changes classification or the shallow overlap read.\nDo not fabricate evidence that does not exist, and do not require a retrospective the requester never\nasked for.\n\n## Entry Route B: Start From Recent-Session Evidence\n\nUse when repeated friction, failures, gaps, or opportunities have been observed in one or more recent\nsessions — session-analysis reports, build handoff reports, validation\u002Fgrading\u002Fcomparison\u002Fsecurity\u002F\ncompleteness\u002Fconsistency findings, user corrections, or durable planning artifacts. If `$ARGUMENTS` names\none of these sources directly (a report path, or one entry from a findings bundle or approved suggestion\nlist an upstream workflow's own human-decision gate already produced), treat it as the evidence source for\nStep 1 below rather than re-deriving it. **This skill classifies and briefs exactly one candidate per\ninvocation** — when a caller's own gate approved several candidates at once (e.g.\n`plugin-lifecycle-maintenance`'s multiSelect finding-selection pick), it invokes this skill once per\ncandidate, never once with the whole list; this skill's own Step 1 never fans a single invocation out into\nmultiple classifications or multiple briefs. Session evidence is a **seed, not the scope** of the concept.\nSee `references\u002Fevidence-routing.md` for the full 6-step\nevidence-handling procedure (identify source, recheck currency, separate symptoms from underlying need,\nmerge duplicates, discard stale\u002Fnon-actionable, obtain explicit approval before promoting to a planned\nchange) and the evidence-source list. If evidence is insufficient, fall back to Entry Route A's focused\ninterview — never require a full retrospective the desired concept doesn't actually depend on.\n\nMost session evidence classifies as Enhance, Repair, or Consolidate rather than Create — those\nclassifications get the full Conception Brief (Steps 4-5 below), since `plugin-lifecycle-maintenance`'s\nexisting workflows have no equivalent deep-definition step of their own.\n\n## Step 1: Normalize Intent and Evidence\n\nBuild a concise problem frame from whichever entry route ran:\n\n- problem or opportunity;\n- target user and concrete scenario;\n- desired outcome;\n- supporting evidence or assumptions;\n- constraints and non-goals;\n- success signals.\n\nLabel every assumption explicitly. Every piece of evidence must identify its source. Never rely on an\nunverified claim that a capability is absent or broken — check the current repository state instead of\ntrusting a stale report.\n\n## Step 2: Classify the Proposed Change\n\n| Classification | Meaning | Likely destination |\n|---|---|---|\n| Create | No suitable existing capability; add a new plugin or component | `plugin-ideation`, then the rest of the upstream pipeline |\n| Enhance | Preserve an existing component and add or improve behavior | `plugin-planning` (if new components are needed) or directly to Fix |\n| Repair | Restore behavior that contradicts the current contract | Directly to Fix — usually skips the full brief (see below) |\n| Consolidate | Multiple components overlap and should be merged or have clearer boundaries | `plugin-planning` plus consistency\u002Factivation review |\n| Reposition | Capability exists but belongs in a different plugin or functional group | `plugin-planning` and marketplace restructuring |\n| Retain | Current behavior is adequate; no implementation is justified | Stop with rationale |\n| Reject\u002FDefer | Benefit, evidence, feasibility, or priority is insufficient | Stop or record a revisit condition |\n\nThe classification may change after Step 3's overlap check — an apparent Create may turn out to be Enhance\nor Consolidate, and, just as often, an apparent Enhance may turn out to be a genuine Create once the\nshallow check shows nothing adjacent actually exists.\n\n**Bypass for narrow repairs:** a Repair classification with an already-known, narrowly-scoped fix and an\nalready-accepted finding does not need the full brief (Steps 4-5) — record the classification and\nevidence, then hand off directly to Fix. This keeps the skill proportional rather than adding ceremony to\na one-line fix.\n\n## Step 3: Check Overlap and Marketplace Fit — Shallow, Classification-Level Only\n\nThis check is deliberately shallow. Its only job is to confirm or revise Step 2's classification — never\nthe exhaustive per-component search `plugin-ideation` runs immediately after a Create handoff, and never\nthe deep evidence-gathering the finding-source tools already ran before this skill started.\n\nInspect, at repository-metadata depth only:\n\n- plugin manifests and plugin purposes;\n- skill, agent, command, hook, and rule descriptions;\n- existing lifecycle ownership;\n- related draft concepts and generated reports.\n\n| Overlap | Meaning |\n|---|---|\n| None | Step 2's classification likely holds; proceed |\n| Partial | Classification may need to shift toward Enhance\u002FConsolidate; reconsider before proceeding |\n| Full | Classification shifts to Retain or Enhance; a Create outcome is very unlikely to survive this finding |\n\n**Do not, here:** activation-trigger overlap analysis, cross-component reference tracing, or\nnaming-candidate validation — escalating past metadata-level inspection is the single most common way this\nskill ends up duplicating `plugin-ideation`. Marketplace placement follows the repository convention: a\nplugin uses `\u003Cdomain>-kit` or `\u003Cdomain>-devkit`, exactly one hyphen immediately before the suffix. For\nCreate outcomes, this skill never proposes a name — that's `plugin-ideation` Step 4's job. For every other\noutcome, record the existing plugin's placement, not a new name.\n\n**If the classification settled here is Retain, Reject, or Defer:** stop now, per Stop Conditions below —\ndo not proceed to Steps 4-7, and do not write a Conception Brief file. State the classification and\nrationale directly in the response instead. This is the only path that produces \"no brief written\" — see\nthe Testing & Validation section's Retain scenario.\n\n## Step 4: Define the Concept — Full Depth Only for Non-Create\n\n**Create:** skip this step entirely. Step 1's problem frame plus the classification and Step 3's shallow\noverlap result is the whole Create handoff payload.\n\n**Enhance \u002F Repair (full-brief) \u002F Consolidate \u002F Reposition:** the accepted concept must specify a\none-sentence value proposition, target users and triggering situations, in-scope capabilities, explicit\nnon-goals, whether the target is a new component or a change to an existing one (never a new plugin — that\nwould be a Create by definition), the proposed marketplace owner and neighboring components,\ndependencies\u002Fconsumers\u002Flifecycle handoffs, compatibility requirements for existing behavior, measurable\nacceptance criteria, principal risks and mitigations, and a baseline contract (behavior that must remain\nunchanged, observed deficiency and evidence, proposed delta, affected callers\u002Fdocs\u002Ftests\u002Factivation\ntriggers, rollback or rejection conditions).\n\n## Step 5: Shape the Implementation Plan\n\n**Create:** skip this step — `plugin-planning` (reached after `plugin-ideation` completes) owns\ncomponent-type\u002Fcount\u002Fdepth decisions.\n\n**Every other non-bypassed classification:** define target files\u002Fcomponent directories, required new or\nmodified components, dependency\u002Fintegration changes, tests or evaluations that demonstrate the concept,\ndocumentation changes, validation\u002Freview\u002Fcommit boundaries, and a recommended execution order. Detailed\nprompts, scripts, schemas, and line-level edits still belong in the subsequent planning\u002Fdesign phase — not\nhere.\n\n## Step 6: Describe the Workflow When the Concept Is Procedural\n\nInclude a workflow only when the proposed capability coordinates three or more dependent states, gates, or\nhandoffs — entry conditions, ordered phases, artifacts consumed\u002Fproduced, human approval gates, stop\u002F\nretry\u002Fresume behavior, downstream owner per handoff, completion criteria. A single focused skill does not\nneed an invented workflow merely to fill this section — mark it not applicable with a one-sentence reason\ninstead.\n\n## Step 7: Decide and Hand Off\n\n**Entry:** the classification is Create, Enhance, Repair (full-brief path), Consolidate, or Reposition — a\nRetain\u002FReject\u002FDefer classification already stopped before reaching this step (see the gate at the end of\nStep 3).\n\nPresent the completed brief via `AskUserQuestion`: approve and proceed \u002F revise \u002F merge with an existing\nconcept \u002F defer with a revisit condition \u002F reject or retain the current state instead. This last pair of\noptions is the user overriding an already-written brief post-hoc (e.g. approving the classification but\ndeciding on reflection the change isn't worth doing) — a different case from the classification-level\nRetain\u002FReject\u002FDefer stop above, which never reaches this step at all. Only an approved concept proceeds —\nthis skill never auto-applies findings or silently turns evidence into a feature.\n\n1. Get a timestamp: `date -u +%Y-%m-%dT%H-%M-%SZ`\n2. Write to `.claude\u002Foutput\u002Fplugin-conception\u002F\u003Cconcept-slug>-\u003Ctimestamp>.md` per\n   `references\u002Fconception-brief-template.md` — for Create, the light variant its own blockquote\n   instructs (the section list lives there, not restated here, to avoid drift); for Enhance\u002FRepair\n   (full-brief)\u002FConsolidate\u002FReposition, all 12 sections in full\n3. Present the artifact link first, then the summary:\n\n```\n📄 Conception Brief written: `.claude\u002Foutput\u002Fplugin-conception\u002F\u003Cconcept-slug>-\u003Ctimestamp>.md`\n```\n\nHand-off target follows the classification:\n\n- **Create →** `plugin-ideation`, with the light brief as its input.\n- **Enhance \u002F Consolidate \u002F Reposition →** `plugin-planning` if new or restructured components are\n  implied, otherwise directly to `plugin-lifecycle-downstream`'s Phase 8 (Consolidated Fix).\n- **Repair (full-brief path) →** directly to `plugin-lifecycle-downstream`'s Phase 8 (Consolidated Fix).\n- **Retain \u002F Reject \u002F Defer →** stop; no downstream hand-off.\n\n**Phase 8 needs a findings bundle, not a brief — this skill never invokes it directly.**\n`plugin-lifecycle-downstream`'s Phase 8 external-entry contract accepts only a schema-validated findings\nbundle (`plugin-rulebook\u002Freferences\u002Fevidence-schema.md`'s Finding\u002FReport-Revision shape, with a\n`baseline_commit` and scope manifest) and refuses anything else — a Conception Brief path does not\nsatisfy it, and this skill has no `Bash(git log:*)` grant to produce `baseline_commit` even if it tried.\nWhere Phase 8 is named as the destination above, the reformatting-and-invoking step belongs to whichever\norchestrator called this skill (e.g. `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin`\nStep 4 already does this conversion) — never to this skill itself.\n\n**Standalone invocation:** for a `plugin-ideation`\u002F`plugin-planning` hand-off, ask via `AskUserQuestion`\nbefore invoking the target — never invoke it silently. For a Phase 8 destination, there is no orchestrator\npresent to do the conversion: state the classification and that a findings-bundle conversion is needed\nbefore Phase 8 can be reached, and stop — never attempt to invoke `plugin-lifecycle-downstream` directly.\n\n**Nested invocation:** when this skill runs as a step inside another orchestrator — `plugin-lifecycle-upstream`'s\nPhase 1 (Conceive), or `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin` Conceive step —\nthe brief-content approval above (approve\u002Frevise\u002Fmerge\u002Fdefer\u002Freject) still runs, but the hand-off\n**invocation** does not: this step ends at \"write the brief and report its classification.\" The calling\norchestrator owns deciding where each classified candidate goes next, since it may be routing several\ncandidates from one run — that decision may be a single consolidated `AskUserQuestion` (as\n`plugin-lifecycle-upstream`'s Gate 1 does) or a deterministic routing rule stated plainly to the user with\nno further ask (as `plugin-lifecycle-maintenance`'s `improve-a-plugin`\u002F`enhance-a-plugin` Step 4 does,\nrouting by classification alone). Either way, never ask twice for the same decision — see those two\ncallers for how each owns this instead.\n\n## Resume Behavior\n\n- A draft brief resumes at the first incomplete section (light briefs have fewer sections to resume\n  through).\n- An approved brief resumes at handoff after the user confirms it is still current.\n- Evidence is rechecked when the underlying target has changed since approval.\n- A legacy Concept Card may still be read directly by `plugin-ideation` without ever passing through this\n  skill — conception is not a mandatory gate in front of an already-existing Concept Card.\n\n## Stop Conditions\n\nStop without implementation when: full overlap makes a new component unnecessary; current behavior\nsatisfies the verified need; evidence is stale or already resolved; the concept lacks a concrete user\noutcome; marketplace ownership cannot be decided without a user choice; risks or cost exceed the accepted\nbenefit; or the user defers or rejects the concept. A clean stop is a valid result and must include its\nrationale — never silently drop a concept without stating why.\n\n## Testing & Validation\n\n**Eval evidence:** `evals\u002Fplugin-conception\u002Fevals.json` — 8 scenarios (Quick Workflow, `workspace\u002Fiteration-1`), 7\u002F8 eval-covered (scenario 1's eval is invalidated by a scripted-premise flaw unrelated to this skill's behavior — see its own `grading.json`); the remaining scenarios below are design-review-verified.\n\n1. **From-scratch Create** — a rough idea with no session evidence; confirm the light brief (per\n   `references\u002Fconception-brief-template.md`'s own light-variant section list) is produced and handed to\n   `plugin-ideation`, never a padded-out full brief\n2. **Session-evidence Enhance** — evidence pointing at friction with an existing component; confirm the\n   full brief is produced (all 12 sections) with a populated baseline contract\n3. **Full overlap** — an idea that duplicates an existing component; confirm the classification shifts to\n   Retain\u002FEnhance and the skill stops with rationale rather than proceeding toward Create\n4. **Stale evidence** — session evidence describing a since-fixed gap; confirm Step 1's current-state\n   reverification catches this and the evidence is discarded, not promoted\n5. **Narrow-repair bypass** — a Repair classification with an already-known fix and an accepted finding;\n   confirm Steps 4-5 are skipped and the skill hands off directly to Fix\n6. **Consolidate** — multiple overlapping components found during Step 3; confirm the full brief is\n   produced and `plugin-planning` plus consistency\u002Factivation review are named as the destination\n7. **Retain\u002Fno-work outcome** — evidence that current behavior is already adequate; confirm a clean stop\n   with stated rationale, no brief written\n8. **Create brief stays light** — construct a Create classification with rich session evidence available;\n   confirm the written brief still marks Scope and Boundaries, Existing-Component Baseline, Implementation\n   Plan, and the \"eventual implementation success\" half of Acceptance Criteria as not applicable, rather\n   than restating what `plugin-ideation`\u002F`plugin-planning` will produce\n\n**Quality gates:**\n- [ ] Step 3's overlap check never exceeds repository-metadata depth — no activation-trigger analysis,\n      cross-component reference tracing, or naming-candidate validation\n- [ ] A Create classification's brief always marks Scope and Boundaries, Existing-Component Baseline,\n      Implementation Plan, and the \"eventual implementation success\" half of Acceptance Criteria not\n      applicable — never a padded-out 12-section document\n- [ ] The narrow-repair bypass (Step 2) is only ever taken when the fix is already known and the finding\n      already accepted — never for an undefined or contested repair\n- [ ] Entry Route B evidence is always rechecked against current repository state before promotion — never\n      copied forward from a source artifact unverified\n- [ ] Step 7's decision is always made via `AskUserQuestion` — this skill never auto-applies a finding or\n      silently promotes evidence into a planned change\n- [ ] **Standalone invocation:** the hand-off target is always invoked via `AskUserQuestion` confirmation\n      first — never invoked silently. **Nested invocation** (inside `plugin-lifecycle-upstream`'s Phase 1\n      or `plugin-lifecycle-maintenance`'s Conceive step): no hand-off is invoked from here at all — only\n      the brief-content approval fires; the calling orchestrator owns the hand-off routing decision (a\n      consolidated ask, or a plainly-stated deterministic rule)\n- [ ] The Conception Brief path is always under `.claude\u002Foutput\u002Fplugin-conception\u002F`\n- [ ] A clean stop (Retain\u002FReject\u002FDefer\u002Ffull overlap\u002Fstale evidence) always states its rationale — never a\n      silent drop\n- [ ] A classification-level Retain\u002FReject\u002FDefer outcome always stops at the end of Step 3, before Steps\n      4-7 run — no Conception Brief file is ever written for it\n\n## Reference Guide\n\n| Resource | Type | Purpose |\n|---|---|---|\n| `scripts\u002Fsmoke_test.py` | In this skill | This skill's own persisted smoke test (frontmatter validity, referenced-file existence, Bash-scope grant consistency) — re-run after any SKILL.md edit |\n| `evals\u002Fplugin-conception\u002F` | Repo root | Persisted `skill-tester` Quick Workflow eval suite (8 scenarios, 7\u002F8 covered) |\n| `references\u002Fconception-brief-template.md` | In this skill | Light and full Conception Brief templates used in Step 7 |\n| `references\u002Fevidence-routing.md` | In this skill | Entry Route B's evidence sources and the 6-step evidence-handling procedure |\n| `plugin-ideation` skill | Sibling skill | Create-classification hand-off target — owns the interview, deep overlap search, and naming |\n| `plugin-planning` skill | Sibling skill | Hand-off target for Enhance\u002FConsolidate\u002FReposition outcomes implying new or restructured components |\n| `plugin-lifecycle-downstream` skill | Sibling skill | Owns Phase 8 (Consolidated Fix) — hand-off target for Repair and no-new-component Enhance outcomes |\n| `plugin-lifecycle-upstream` skill | Sibling skill | Runs this skill as Phase 1 (Conceive) |\n| `plugin-lifecycle-maintenance` skill | Sibling skill | Runs this skill as a step inside `improve-a-plugin`\u002F`enhance-a-plugin`, after the human finding-selection gate |\n",{"name":5,"description":7,"allowed-tools":23,"argument-hint":24},"Read Glob Grep AskUserQuestion Write Bash(date:*) Skill","[rough idea, problem statement, or session-evidence source]","main","1bf12900cac6f28e346abd2912bdb78d306de723","plugins\u002Fplugin-devkit\u002Fskills\u002Fplugin-conception","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace","https:\u002F\u002Fgithub.com\u002FAndreHahm","2026-11-15 04:26:16",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add AndreHahm\u002Fandres-cc-marketplace --skill plugin-conception",{"tool":37,"command":38},"bunx","bunx skills add AndreHahm\u002Fandres-cc-marketplace --skill plugin-conception",{"tool":40,"command":41},"pnpm","pnpm dlx skills add AndreHahm\u002Fandres-cc-marketplace --skill plugin-conception","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-devkit\u002Fskills\u002Fplugin-conception\u002FSKILL.md",true,[45,51,55],{"path":46,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":49,"sourceHtmlUrl":50},"references\u002Fconception-brief-template.md","REFERENCE","参考文档",12070,"https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-devkit\u002Fskills\u002Fplugin-conception\u002Freferences\u002Fconception-brief-template.md",{"path":52,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":53,"sourceHtmlUrl":54},"references\u002Fevidence-routing.md",4275,"https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-devkit\u002Fskills\u002Fplugin-conception\u002Freferences\u002Fevidence-routing.md",{"path":56,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":59,"sourceHtmlUrl":60},"scripts\u002Fsmoke_test.py","SCRIPT","脚本",3942,"https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-devkit\u002Fskills\u002Fplugin-conception\u002Fscripts\u002Fsmoke_test.py",[62,69,76,83,90,97],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":18},"01M79R0K918Z0XDM74R0EY5H3P","demo","Fixture skill used by scripts.marketplace_ci sync tests.","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Ftests\u002Fmarketplace_ci\u002Ffixtures\u002Fplugins\u002Fsample-kit\u002Fskills\u002Fdemo\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":18},"01M79R08VMDJA2FPJKH3ZM309K","export-demo","Fixture skill used by scripts.marketplace_ci export tests.","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Ftests\u002Fmarketplace_ci\u002Ffixtures\u002Fclaude\u002Fskills\u002Fexport-demo\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":18},"01M79R042FS581DS2JNW5AZ1MK","work-to-development","Validate an accepted Linear Issue's readiness for implementation (criteria, dependencies, duplicate\u002Fexisting artifacts) and request governed branch\u002Fworktree creation through git-kit:starting-work — recording work-started Git\u002FGitHub evidence and a deliberate Linear Started transition only after Git identity is read back and verified. Use when asked to start work on a Linear issue, begin development for an accepted issue, or prepare an accepted issue for implementation. Never creates a branch directly — always delegates to git-kit.","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fworkmanagement-kit\u002Fskills\u002Fwork-to-development\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":18},"01M79QZZD4EHB82P72P4N92BKT","work-linking","Create and maintain stable cross-system links and authority labels between Notion and Linear records, and detect\u002Frepair drift between them. Use when asked to link a Notion record to a Linear record, check for drift between Notion and Linear, or repair a broken\u002Fstale link. Repairs only the bounded non-authoritative copy or link — never chooses by newest timestamp, and never creates a reverse-write loop against either system's own authority.","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fworkmanagement-kit\u002Fskills\u002Fwork-linking\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":18},"01M79QZVFA3GTVNZMJ51NESVSP","status-and-learning","Produce a dated, explicitly non-live Linear-to-Notion progress summary, or capture an outcome\u002Flearning after work completes. Use when asked to summarize progress to Notion, report portfolio status, or capture what was learned from a completed piece of work — not to disposition open follow-up items (see open-item-management for that). Never runs automatically or continuously — every summary is a deliberate, approved snapshot, not a live sync.","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fworkmanagement-kit\u002Fskills\u002Fstatus-and-learning\u002FSKILL.md",[],[],{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":18},"01M79QZNSKSABK7X2PT93X8QT5","repository-gates","Discover, delegate, record, and invalidate this repository's actual Git\u002FGitHub lifecycle gates (pre-commit, pre-push, PR required checks, Codex delta review, review requirements, merge rights) and the repository-policy provider profile that maps governed operations to git-kit. Use when asked what gates apply to this repository, to discover repository policy before starting Wave 2 work, or to check whether prior gate evidence is still valid after new commits. Discovery and delegation only — never executes a gate's own internal logic, and never invents a gate the repository doesn't actually define (in particular, never assumes a \"Review Changes\" gate exists).","https:\u002F\u002Fgithub.com\u002FAndreHahm\u002Fandres-cc-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fworkmanagement-kit\u002Fskills\u002Frepository-gates\u002FSKILL.md",[],[]]