[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77JZP5RXGA3FHQZ868CS8HP":3,"skill-same:01M77JZP5RXGA3FHQZ868CS8HP":45},{"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":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M77JZP5RXGA3FHQZ868CS8HP","calibration-loop",null,"Converge on an ambiguous preference through concrete either\u002For instances instead of re-asking the same abstract question. Use when brainstorm divergence is unsure, a doc-review disposition is disputed, or feedback scope classification is ambiguous. Does not resolve unrelated ambiguity or bypass human confirmation on the converged rule.","grdavies\u002Fshipwright","grdavies","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F2638082?v=4","01M77JZCZJTWB8WBWX4NNW8MNP","Ross Davies","01M77JZD0S2Q4JZEM0310T9A73",2,"MIT","https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fcalibration-loop\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: calibration-loop\ndescription: Converge on an ambiguous preference through concrete either\u002For instances instead of re-asking the same abstract question. Use when brainstorm divergence is unsure, a doc-review disposition is disputed, or feedback scope classification is ambiguous. Does not resolve unrelated ambiguity or bypass human confirmation on the converged rule.\n---\n# Calibration loop\n\nA reusable convergence primitive for genuine either\u002For ambiguity — not a general Q&A loop. Presents one\nconcrete instance per turn, restates the inferred principle, and stops on stability.\n\n**Model tier:** cheap — resolve via `python3 scripts\u002Fsw_bootstrap.py resolve-model-tier.py -- --skill calibration-loop`.\n\n## When to use\n\nOnly when a prior attempt to resolve a preference\u002Fdisposition abstractly has already failed (the operator\nsaid \"unsure\", two reviewers disagree, or scope classification is genuinely ambiguous) — never as a first\nresort over a direct question.\n\n## Fixed verdict set\n\nEvery turn presents one **concrete** either\u002For instance (a specific example, not the abstract question again)\nand accepts exactly one of a fixed vocabulary:\n\n| Verdict | Meaning |\n| --- | --- |\n| `A` | This instance resolves toward option A |\n| `B` | This instance resolves toward option B |\n| `either` | Genuinely indifferent on this instance (informative — narrows the principle to \"doesn't matter here\") |\n| `neither` | Instance is malformed or off-tension — discard, generate a better one |\n| `more-info` | Operator needs one clarifying fact before verdicting this instance — answer, then re-present the same instance |\n\nNever accept free text as the primary channel — free text is allowed only as an annotation alongside one of\nthe five verdicts above, never as a substitute for one.\n\n## Procedure\n\n1. **Frame the tension** — restate the two poles (A vs B) in one sentence from the blocked artifact\n   (divergence stance pair, disputed disposition, ambiguous-scope routing).\n2. **Present one concrete instance** — a specific, realistic case that would force a choice between A and B.\n   Never present the abstract question a second time.\n3. **Accept a verdict** from the fixed set above.\n4. **Restate the inferred principle** — after each verdict, state the working rule in one sentence (\"so:\n   when X, prefer A; when Y, prefer B\") and ask for a quick confirm\u002Fcorrect on the restatement, not a new\n   open question.\n5. **Adapt toward unexplored seams** — the next instance MUST probe a boundary the current principle has not\n   yet been tested against (not a near-duplicate of the last instance). Track covered vs. unexplored seams\n   explicitly in scratch notes for the run.\n6. **Stop on stability** — converge when either:\n   - two consecutive instances confirm the same restated principle with no correction, or\n   - a hard cap of 5 instances is reached (whichever first). On cap-out without stability, record\n     `converged: false` and escalate to explicit human decision — never silently pick a side.\n7. **Write the converged rule into the blocked artifact** — the calling consumer (brainstorm, doc-review,\n   feedback) persists the final principle at its own persistence point (see **Consumers** below); this skill\n   never writes on its own — it returns the converged principle + verdict trail to the caller.\n\n## Consumers (wired now)\n\n| Consumer | Trigger | Where the converged rule lands |\n| --- | --- | --- |\n| `\u002Fsw-brainstorm` divergence (unsure routing) | Operator \"unsure\" between ~2 viable stances | Requirements doc **Key Decisions** — chosen stance + converged principle |\n| `\u002Fsw-doc-review` disposition disputes | Synthesizer\u002Freviewer disagreement on `safe_auto` vs `gated_auto`\u002F`manual` for a finding class | Review-round synthesis notes; principle informs the disposition for that finding and future same-class findings in the run |\n| `\u002Fsw-feedback` ambiguous-scope classification | Signal cannot be cleanly classed `gap-capture` vs `brainstorm` (or `substantial` vs `trivial`) | Route record (`skills\u002Ffeedback\u002Freferences\u002Froute-record.md`) — converged principle recorded alongside the route decision |\n\n## Guardrails\n\n- Never a substitute for a direct, answerable question — only for genuine either\u002For ambiguity after a first\n  abstract attempt.\n- Never accept the same instance twice in a row as \"resolved\" — each turn must probe a new seam.\n- Cap-out (`converged: false`) always escalates to explicit human decision; never auto-resolves silently.\n- This skill informs the blocked artifact — it does not itself freeze, merge, or dispatch anything.\n","# Calibration loop\n\nA reusable convergence primitive for genuine either\u002For ambiguity — not a general Q&A loop. Presents one\nconcrete instance per turn, restates the inferred principle, and stops on stability.\n\n**Model tier:** cheap — resolve via `python3 scripts\u002Fsw_bootstrap.py resolve-model-tier.py -- --skill calibration-loop`.\n\n## When to use\n\nOnly when a prior attempt to resolve a preference\u002Fdisposition abstractly has already failed (the operator\nsaid \"unsure\", two reviewers disagree, or scope classification is genuinely ambiguous) — never as a first\nresort over a direct question.\n\n## Fixed verdict set\n\nEvery turn presents one **concrete** either\u002For instance (a specific example, not the abstract question again)\nand accepts exactly one of a fixed vocabulary:\n\n| Verdict | Meaning |\n| --- | --- |\n| `A` | This instance resolves toward option A |\n| `B` | This instance resolves toward option B |\n| `either` | Genuinely indifferent on this instance (informative — narrows the principle to \"doesn't matter here\") |\n| `neither` | Instance is malformed or off-tension — discard, generate a better one |\n| `more-info` | Operator needs one clarifying fact before verdicting this instance — answer, then re-present the same instance |\n\nNever accept free text as the primary channel — free text is allowed only as an annotation alongside one of\nthe five verdicts above, never as a substitute for one.\n\n## Procedure\n\n1. **Frame the tension** — restate the two poles (A vs B) in one sentence from the blocked artifact\n   (divergence stance pair, disputed disposition, ambiguous-scope routing).\n2. **Present one concrete instance** — a specific, realistic case that would force a choice between A and B.\n   Never present the abstract question a second time.\n3. **Accept a verdict** from the fixed set above.\n4. **Restate the inferred principle** — after each verdict, state the working rule in one sentence (\"so:\n   when X, prefer A; when Y, prefer B\") and ask for a quick confirm\u002Fcorrect on the restatement, not a new\n   open question.\n5. **Adapt toward unexplored seams** — the next instance MUST probe a boundary the current principle has not\n   yet been tested against (not a near-duplicate of the last instance). Track covered vs. unexplored seams\n   explicitly in scratch notes for the run.\n6. **Stop on stability** — converge when either:\n   - two consecutive instances confirm the same restated principle with no correction, or\n   - a hard cap of 5 instances is reached (whichever first). On cap-out without stability, record\n     `converged: false` and escalate to explicit human decision — never silently pick a side.\n7. **Write the converged rule into the blocked artifact** — the calling consumer (brainstorm, doc-review,\n   feedback) persists the final principle at its own persistence point (see **Consumers** below); this skill\n   never writes on its own — it returns the converged principle + verdict trail to the caller.\n\n## Consumers (wired now)\n\n| Consumer | Trigger | Where the converged rule lands |\n| --- | --- | --- |\n| `\u002Fsw-brainstorm` divergence (unsure routing) | Operator \"unsure\" between ~2 viable stances | Requirements doc **Key Decisions** — chosen stance + converged principle |\n| `\u002Fsw-doc-review` disposition disputes | Synthesizer\u002Freviewer disagreement on `safe_auto` vs `gated_auto`\u002F`manual` for a finding class | Review-round synthesis notes; principle informs the disposition for that finding and future same-class findings in the run |\n| `\u002Fsw-feedback` ambiguous-scope classification | Signal cannot be cleanly classed `gap-capture` vs `brainstorm` (or `substantial` vs `trivial`) | Route record (`skills\u002Ffeedback\u002Freferences\u002Froute-record.md`) — converged principle recorded alongside the route decision |\n\n## Guardrails\n\n- Never a substitute for a direct, answerable question — only for genuine either\u002For ambiguity after a first\n  abstract attempt.\n- Never accept the same instance twice in a row as \"resolved\" — each turn must probe a new seam.\n- Cap-out (`converged: false`) always escalates to explicit human decision; never auto-resolves silently.\n- This skill informs the blocked artifact — it does not itself freeze, merge, or dispatch anything.\n",{"name":5,"description":7},"main","2633b37cdc50086084cc9c0ce6c994871d3e7aad","core\u002Fskills\u002Fcalibration-loop",0,"https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright","https:\u002F\u002Fgithub.com\u002Fgrdavies","2026-11-03 23:44:15",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add grdavies\u002Fshipwright --skill calibration-loop",{"tool":37,"command":38},"bunx","bunx skills add grdavies\u002Fshipwright --skill calibration-loop",{"tool":40,"command":41},"pnpm","pnpm dlx skills add grdavies\u002Fshipwright --skill calibration-loop","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fcalibration-loop\u002FSKILL.md",true,[],[46,53,60,67,73,80],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":19},"01M77K4Z004VAMY705XQKJAH4Y","wrong-name","Missing use when phrase and wrong directory name for critical lint failure.","https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fscripts\u002Ftest\u002Ffixtures\u002Fworkflow-pack-sdk\u002Fnegative\u002Fbad-instruction\u002Fskills\u002Fbad-skill\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M77K4WR3WP22SBT6053PYWAK","demo-skill","Demo workflow pack skill fixture. Use when validating workflow-pack-sdk golden paths in CI.","https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fscripts\u002Ftest\u002Ffixtures\u002Fworkflow-pack-sdk\u002Fgolden\u002Fvalid-pack\u002Fskills\u002Fdemo-skill\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M77K4SV48XDDDQ81PD09G9N6","good-skill","Validates Agent Skills spec conformance for fixture trees. Use when testing skills-spec-guard pass paths across core and dist mirrors. Does not ship production skills.","https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fscripts\u002Ftest\u002Ffixtures\u002Fskills-spec\u002Fpass-valid\u002Fcore\u002Fskills\u002Fgood-skill\u002FSKILL.md",[],[],{"skillNo":68,"name":48,"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":19},"01M77K4PWVT96VMTMMFE2Y80GJ","Fixture with name\u002Fdir mismatch for guard failure tests. Use when verifying skills-spec-guard rejects invalid identity. Does not ship production skills.","https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fscripts\u002Ftest\u002Ffixtures\u002Fskills-spec\u002Ffail-name-mismatch\u002Fcore\u002Fskills\u002Fgood-skill\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M77K4KSS5Z0Q0DEAA8PV9316","bad-description","Missing the required when trigger phrase in this fixture skill.","https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fscripts\u002Ftest\u002Ffixtures\u002Fskills-spec\u002Ffail-description-shape\u002Fcore\u002Fskills\u002Fbad-description\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M77K4H9HK5FDF84D65228XFA","dropin-demo","https:\u002F\u002Fgithub.com\u002Fgrdavies\u002Fshipwright\u002Fblob\u002Fmain\u002Fscripts\u002Ftest\u002Ffixtures\u002Fcapability-select\u002Fdropin\u002Fcore\u002Fskills\u002Fdropin-demo\u002FSKILL.md",[],[]]