[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M72WKWGWE149Y5NBPC1ZR4AC":3,"skill-same:01M72WKWGWE149Y5NBPC1ZR4AC":74},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"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":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":26,"repositoryHomepageUrl":32,"ownerHtmlUrl":33,"ownerBlogUrl":32,"ownerDescription":34,"updatedAt":35,"install":36,"resources":49},"01M72WKWGWE149Y5NBPC1ZR4AC","idea",null,"Runs a structured idea-convergence session on a forge issue: opens the session, triages and compares competing proposals, tallies votes under the solo, unanimity or vote governance mode, and closes with one or more SPECS issues or a no-consensus record. Use when several competing, overlapping or complementary proposals address the same parent issue and must converge before a spec is written, or when asked to open, tally or close an IDEA session. Trigger terms: IDEA session, converge proposals, competing designs, candidate, VOTE comment, quorum, unanimity, no consensus. Not for a ticket whose intent is already clear (spec-author), reviewing a pull request (pr-reviewer) or open-ended brainstorming.","crewrig\u002Fcrewrig","crewrig","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F283580446?v=4","01M72WDBD9H0YW5JBHNZ0TEVWB","01M72WDBEBNRBAS94AXRPEVYNK",24,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fidea\u002FSKILL.md",[],[],"2026-11-10 22:39:42",false,"---\nname: idea\ndescription: \"Runs a structured idea-convergence session on a forge issue: opens the session, triages and compares competing proposals, tallies votes under the solo, unanimity or vote governance mode, and closes with one or more SPECS issues or a no-consensus record. Use when several competing, overlapping or complementary proposals address the same parent issue and must converge before a spec is written, or when asked to open, tally or close an IDEA session. Trigger terms: IDEA session, converge proposals, competing designs, candidate, VOTE comment, quorum, unanimity, no consensus. Not for a ticket whose intent is already clear (spec-author), reviewing a pull request (pr-reviewer) or open-ended brainstorming.\"\nlicense: Apache-2.0\ncompatibility: \"Requires Node.js 24 (scripts run as node scripts\u002F\u003Cname>.ts) and the forge CLI (gh for GitHub; glab or tea users export the issue JSON and translate the printed commands).\"\nmetadata:\n  provenance:\n    canonical: \"https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\"\n    feedback: \"https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\"\n    version: \"1.2.0\"\n---\n\n\n# Idea\n\nAn IDEA session settles which of several proposals becomes the spec's intent, and leaves an auditable record on a forge issue: proposals, triage, comparison, votes and outcome are all comments on one session issue. It sits before SPECS and is optional. When a single clear intent exists, go straight to `spec-author`.\n\nMechanical steps are scripts; you supply the judgement (clustering ideas, comparing, composing, designating the candidate). Run each as `node \u003Cskill-dir>\u002Fscripts\u002F\u003Cname>.ts`. All take `--json` for `{ok, data, findings}` and `--help`; exit `2` means a wrong invocation and `3` a failed environment (usually `gh` unavailable or unauthenticated), never a verdict on the session.\n\n| When | Command | Exit `1` means |\n|---|---|---|\n| Opening | `idea-open.ts --mode \u003Cm> --issue \u003CN> --description \u003Ct>` plus the governance options of the mode | Never used. `--dry-run` prints the `gh` commands and the issue body and runs nothing. |\n| Before any decision to close | `idea-tally.ts --issue \u003CN>` (or `--comments-file \u003Cexported.json>`) | No usable candidate: `candidate-required` (two or more proposals survive and none is designated) or `no-proposals`. Designate a candidate first. |\n| Closing | `idea-close.ts --issue \u003CS> --outcome resolved\\|no-consensus --parent \u003CN> ...` | Refused: `threshold-not-met` or `winner-mismatch`. Do not close as resolved. |\n\nRead `data` of the tally, not only the exit code: `threshold_met` is the verdict, `missing` names who has not voted, `ignored` and the warning findings say which comments were not counted and why. Pass the tally's `--json` output to `idea-close.ts --tally-file` so the closing comment carries the real counts.\n\n## Governance modes\n\nThe mode and its parameters are fixed when the session issue is created, and are written into its body by `idea-open.ts` (the tally reads them back from there). Changing them mid-session would let the outcome be steered after the votes are visible, so the only recourse is to close the session as no-consensus and open a new one.\n\n| Mode | Parameters | Threshold |\n|---|---|---|\n| `solo` | `--owner \u003Clogin>`: one decision authority. Default to the invoking user's login when the user gives none. | The owner posts `VOTE: APPROVE`. A `REJECT` is a veto that leaves the session open. |\n| `unanimity` | `--owners \u003Ca,b,...>` | Every owner has approved. One `REJECT` blocks, and the session stays open. |\n| `vote` | `--voters \u003Ca,b,...>` and `--quorum \u003Cpct>` (1-100) | Approvals are at least `quorum` percent of the approve and reject votes, with at least one such vote. `ABSTAIN` is left out of the denominator. |\n\nOptional `--dimensions \"a;b\"` replaces the default comparison dimensions at open time; they cannot change later.\n\n## Session workflow\n\n1. **Open.** Run `idea-open.ts`. It creates `[IDEA] \u003Cdescription> (parent #N)` with the labels `idea-session` and `idea-\u003Cmode>` and comments on the parent issue. On a forge other than GitHub, run it with `--dry-run` and translate the commands.\n2. **Collect proposals** as comments on the session issue, each in this shape, so they compare cleanly:\n\n   ```markdown\n   ## Proposal: \u003Ctitle>\n\n   **Submitter:** \u003Cidentity>\n\n   ### Summary\n   \u003Cone to three sentences>\n\n   ### Rationale\n   \u003Cwhy this addresses the parent issue>\n\n   ### Trade-offs\n   \u003Cadvantages and costs, explicit>\n\n   ### Open questions\n   \u003Cunresolved points that may affect feasibility or design>\n   ```\n\n3. **Triage.** Judge each proposal on relevance (does it address the parent issue), feasibility (realizable within reasonable constraints) and non-duplication (a distinct mechanism from an accepted proposal, not merely different wording; the earlier submission is the one kept). Reject with a comment whose first line is exactly `[TRIAGE-REJECTED] Proposal: \u003Ctitle>`, followed by `**Criterion failed:**` and a `**Reason:**` grounded in the proposal. Rejected proposals take no further part.\n4. **Compare.** Post one comment presenting every survivor against every dimension. Defaults: implementation complexity, architectural alignment (existing architecture, ADRs, constraints), reversibility.\n5. **Compose (optional).** When survivors are complementary rather than exclusive, post a hybrid as `[COMPOSITION] Proposal: \u003Ctitle>` naming its sources (`**Composed from:**`), in the standard proposal format. It then competes like any survivor.\n6. **Designate the candidate.** With two or more survivors, post a top-level comment whose first line is exactly `[CANDIDATE] Proposal: \u003Ctitle>`, naming a surviving proposal or composition. Votes bind to the latest valid designation, and a new valid designation voids every earlier vote because voters approved something else; they must recast. A designation naming an unknown title designates nothing and voids nothing. With exactly one survivor and no designation ever posted, that survivor is the implicit candidate.\n7. **Vote.** Eligible identities post a top-level comment that is exactly `VOTE: APPROVE`, `VOTE: REJECT` or `VOTE: ABSTAIN` (reactions and replies do not count; a voter's latest vote stands). Run `idea-tally.ts` as often as useful.\n8. **Close.** Close only on a tally with `threshold_met: true`, or when the owner decides to stop.\n   - Resolved: create the SPECS issues first, each seeded with the winner's summary as its intent and linking the session issue; then run `idea-close.ts --outcome resolved --spec \u003CN:description> ... --tally-file \u003Ctally.json>`.\n   - No consensus (allowed at any time, no SPECS issue is created, the parent stays open): `idea-close.ts --outcome no-consensus --reason \"\u003Cwhy>\"`. The reason is the useful part of the record; state what blocked convergence.\n\n   Every closed session carries exactly one terminal label (`idea-resolved` or `idea-no-consensus`); the script adds it, so do not close a session by hand.\n\n## Retroactive review loop\n\nThe session is the audit record of how an intent was chosen. A review finding of class `spec` that traces to an inadequately converged set of proposals may justify a new session on the parent issue; the orchestrator decides, it is not automatic. Finding classes: `references\u002Fshared-finding-class-taxonomy.md`.\n\n## Friction\n\nReport workflow friction as described in `references\u002Fshared-friction-reporting.md`.\n","# Idea\n\nAn IDEA session settles which of several proposals becomes the spec's intent, and leaves an auditable record on a forge issue: proposals, triage, comparison, votes and outcome are all comments on one session issue. It sits before SPECS and is optional. When a single clear intent exists, go straight to `spec-author`.\n\nMechanical steps are scripts; you supply the judgement (clustering ideas, comparing, composing, designating the candidate). Run each as `node \u003Cskill-dir>\u002Fscripts\u002F\u003Cname>.ts`. All take `--json` for `{ok, data, findings}` and `--help`; exit `2` means a wrong invocation and `3` a failed environment (usually `gh` unavailable or unauthenticated), never a verdict on the session.\n\n| When | Command | Exit `1` means |\n|---|---|---|\n| Opening | `idea-open.ts --mode \u003Cm> --issue \u003CN> --description \u003Ct>` plus the governance options of the mode | Never used. `--dry-run` prints the `gh` commands and the issue body and runs nothing. |\n| Before any decision to close | `idea-tally.ts --issue \u003CN>` (or `--comments-file \u003Cexported.json>`) | No usable candidate: `candidate-required` (two or more proposals survive and none is designated) or `no-proposals`. Designate a candidate first. |\n| Closing | `idea-close.ts --issue \u003CS> --outcome resolved\\|no-consensus --parent \u003CN> ...` | Refused: `threshold-not-met` or `winner-mismatch`. Do not close as resolved. |\n\nRead `data` of the tally, not only the exit code: `threshold_met` is the verdict, `missing` names who has not voted, `ignored` and the warning findings say which comments were not counted and why. Pass the tally's `--json` output to `idea-close.ts --tally-file` so the closing comment carries the real counts.\n\n## Governance modes\n\nThe mode and its parameters are fixed when the session issue is created, and are written into its body by `idea-open.ts` (the tally reads them back from there). Changing them mid-session would let the outcome be steered after the votes are visible, so the only recourse is to close the session as no-consensus and open a new one.\n\n| Mode | Parameters | Threshold |\n|---|---|---|\n| `solo` | `--owner \u003Clogin>`: one decision authority. Default to the invoking user's login when the user gives none. | The owner posts `VOTE: APPROVE`. A `REJECT` is a veto that leaves the session open. |\n| `unanimity` | `--owners \u003Ca,b,...>` | Every owner has approved. One `REJECT` blocks, and the session stays open. |\n| `vote` | `--voters \u003Ca,b,...>` and `--quorum \u003Cpct>` (1-100) | Approvals are at least `quorum` percent of the approve and reject votes, with at least one such vote. `ABSTAIN` is left out of the denominator. |\n\nOptional `--dimensions \"a;b\"` replaces the default comparison dimensions at open time; they cannot change later.\n\n## Session workflow\n\n1. **Open.** Run `idea-open.ts`. It creates `[IDEA] \u003Cdescription> (parent #N)` with the labels `idea-session` and `idea-\u003Cmode>` and comments on the parent issue. On a forge other than GitHub, run it with `--dry-run` and translate the commands.\n2. **Collect proposals** as comments on the session issue, each in this shape, so they compare cleanly:\n\n   ```markdown\n   ## Proposal: \u003Ctitle>\n\n   **Submitter:** \u003Cidentity>\n\n   ### Summary\n   \u003Cone to three sentences>\n\n   ### Rationale\n   \u003Cwhy this addresses the parent issue>\n\n   ### Trade-offs\n   \u003Cadvantages and costs, explicit>\n\n   ### Open questions\n   \u003Cunresolved points that may affect feasibility or design>\n   ```\n\n3. **Triage.** Judge each proposal on relevance (does it address the parent issue), feasibility (realizable within reasonable constraints) and non-duplication (a distinct mechanism from an accepted proposal, not merely different wording; the earlier submission is the one kept). Reject with a comment whose first line is exactly `[TRIAGE-REJECTED] Proposal: \u003Ctitle>`, followed by `**Criterion failed:**` and a `**Reason:**` grounded in the proposal. Rejected proposals take no further part.\n4. **Compare.** Post one comment presenting every survivor against every dimension. Defaults: implementation complexity, architectural alignment (existing architecture, ADRs, constraints), reversibility.\n5. **Compose (optional).** When survivors are complementary rather than exclusive, post a hybrid as `[COMPOSITION] Proposal: \u003Ctitle>` naming its sources (`**Composed from:**`), in the standard proposal format. It then competes like any survivor.\n6. **Designate the candidate.** With two or more survivors, post a top-level comment whose first line is exactly `[CANDIDATE] Proposal: \u003Ctitle>`, naming a surviving proposal or composition. Votes bind to the latest valid designation, and a new valid designation voids every earlier vote because voters approved something else; they must recast. A designation naming an unknown title designates nothing and voids nothing. With exactly one survivor and no designation ever posted, that survivor is the implicit candidate.\n7. **Vote.** Eligible identities post a top-level comment that is exactly `VOTE: APPROVE`, `VOTE: REJECT` or `VOTE: ABSTAIN` (reactions and replies do not count; a voter's latest vote stands). Run `idea-tally.ts` as often as useful.\n8. **Close.** Close only on a tally with `threshold_met: true`, or when the owner decides to stop.\n   - Resolved: create the SPECS issues first, each seeded with the winner's summary as its intent and linking the session issue; then run `idea-close.ts --outcome resolved --spec \u003CN:description> ... --tally-file \u003Ctally.json>`.\n   - No consensus (allowed at any time, no SPECS issue is created, the parent stays open): `idea-close.ts --outcome no-consensus --reason \"\u003Cwhy>\"`. The reason is the useful part of the record; state what blocked convergence.\n\n   Every closed session carries exactly one terminal label (`idea-resolved` or `idea-no-consensus`); the script adds it, so do not close a session by hand.\n\n## Retroactive review loop\n\nThe session is the audit record of how an intent was chosen. A review finding of class `spec` that traces to an inadequately converged set of proposals may justify a new session on the parent issue; the orchestrator decides, it is not automatic. Finding classes: `references\u002Fshared-finding-class-taxonomy.md`.\n\n## Friction\n\nReport workflow friction as described in `references\u002Fshared-friction-reporting.md`.\n",{"name":5,"license":14,"metadata":23,"description":7,"compatibility":27},{"provenance":24},{"version":25,"feedback":26,"canonical":26},"1.2.0","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig","Requires Node.js 24 (scripts run as node scripts\u002F\u003Cname>.ts) and the forge CLI (gh for GitHub; glab or tea users export the issue JSON and translate the printed commands).","main","6d7f2262eb86a02d21a60a8147ed6f8a27d7b8eb",".github\u002Fskills\u002Fidea",6,"https:\u002F\u002Fcrewrig.org","https:\u002F\u002Fgithub.com\u002Fcrewrig","This is the organisation behind the crewrig framework","2026-11-15 04:46:08",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add crewrig\u002Fcrewrig --skill idea",{"tool":42,"command":43},"bunx","bunx skills add crewrig\u002Fcrewrig --skill idea",{"tool":45,"command":46},"pnpm","pnpm dlx skills add crewrig\u002Fcrewrig --skill idea","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fidea\u002FSKILL.md",true,[50,56,60,66,70],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"references\u002Fshared-finding-class-taxonomy.md","REFERENCE","参考文档",1287,"https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fidea\u002Freferences\u002Fshared-finding-class-taxonomy.md",{"path":57,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002Fshared-friction-reporting.md",769,"https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fidea\u002Freferences\u002Fshared-friction-reporting.md",{"path":61,"resourceType":62,"resourceTypeDesc":63,"sizeBytes":64,"sourceHtmlUrl":65},"scripts\u002Fidea-close.ts","SCRIPT","脚本",9252,"https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fidea\u002Fscripts\u002Fidea-close.ts",{"path":67,"resourceType":62,"resourceTypeDesc":63,"sizeBytes":68,"sourceHtmlUrl":69},"scripts\u002Fidea-open.ts",8064,"https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fidea\u002Fscripts\u002Fidea-open.ts",{"path":71,"resourceType":62,"resourceTypeDesc":63,"sizeBytes":72,"sourceHtmlUrl":73},"scripts\u002Fidea-tally.ts",13140,"https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fidea\u002Fscripts\u002Fidea-tally.ts",[75,83,88,93,98,103],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":82},"01M76AJKRKGGYZFJJHKE32Q5ZZ","fixture","Fixture skill. Mentions scripts\u002Ffrontmatter-only.ts which must be ignored.","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Ffixtures\u002Fskill-scripts-citations\u002Fyaml-example\u002FSKILL.md",[],[],"2026-11-12 13:55:43",{"skillNo":84,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":82},"01M76AJJHS0AHW5HJJQCVVJ4DH","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Ffixtures\u002Fskill-scripts-citations\u002Funterminated\u002FSKILL.md",[],[],{"skillNo":89,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":82},"01M76AJH8VJ2NAE000285J1D8F","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Ffixtures\u002Fskill-scripts-citations\u002Funresolved\u002FSKILL.md",[],[],{"skillNo":94,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":95,"categories":96,"keywords":97,"publishedAt":82},"01M76AJG4M862QH00H5QVVT60P","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Ffixtures\u002Fskill-scripts-citations\u002Fstray\u002FSKILL.md",[],[],{"skillNo":99,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":100,"categories":101,"keywords":102,"publishedAt":82},"01M76AJEZKZVMWZXCKXCFNQQ6B","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Ffixtures\u002Fskill-scripts-citations\u002Fresolving\u002FSKILL.md",[],[],{"skillNo":104,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":82},"01M76AJD4G4K2YK2SMN8NSEG9R","https:\u002F\u002Fgithub.com\u002Fcrewrig\u002Fcrewrig\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Ffixtures\u002Fskill-scripts-citations\u002Freferences-unfenced\u002FSKILL.md",[],[]]