[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6R5WBT4JJA032YSX1TYEJTS":3,"skill-same:01M6R5WBT4JJA032YSX1TYEJTS":103},{"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":30,"linkOnly":31,"markdownContent":32,"markdownBody":33,"frontMatter":34,"refName":35,"commitSha":36,"skillDirectory":37,"forksCount":13,"repositoryHtmlUrl":38,"repositoryHomepageUrl":6,"ownerHtmlUrl":39,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":40,"install":41,"resources":54},"01M6R5WBT4JJA032YSX1TYEJTS","northstar",null,"Planning-only intake: turn intent into a tracked, sliced plan and A→B handoff; never implement product changes. Use before autobahn execution.","r3dlex\u002Fskills","r3dlex","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F291373160?v=4","01M6PEFNGGBVB7SHAPWXFFX933","01M6R5SDWFHT94734KJFMYE1SX",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002FSKILL.md",[],[5,18,19,20,21,22,23,24,25,26,27,28,29],"handoff","implement","planning","product","intake","execution","plan","use","never","only","intent","changes","2026-11-10 10:12:30",false,"---\nname: northstar\ndescription: \"Planning-only intake: turn intent into a tracked, sliced plan and A→B handoff; never implement product changes. Use before autobahn execution.\"\neval: northstar\n---\n\n# Northstar\n\nNorthstar converges intent into a tracked, sliced plan inside a repo initialized\nwith `ai-catapult-init` v3. This lightweight **composer** delegates to existing skills; its output is the A→B handoff that `autobahn` consumes.\n\n## Execution boundary (hard stop)\n\nNorthstar is **planning-only**. Never implement a sliced goal, modify product\ncode or tests, run implementation engines, or start `autobahn` while this skill\nis active. Writes are limited to the planning\u002Ftracking artifacts required below.\nAfter verifying the A→B handoff, **stop and report it**; implementation requires\na separate, explicit `autobahn` invocation. Never continue into implementation\nin the same run, even when the initiating prompt also asks for implementation.\n\n## Quick Start\n\n1. Run the prereq gate (fail-closed) against the repo root:\n   `bash northstar\u002Fprereq-check.sh --root .`\n2. Run the interview loop, then always raise an issue.\n3. Run `ralplan` to produce sliced goals.\n4. Publish the reviewed versioned bundle: `bash northstar\u002Fhandoff-write.sh --root \u003Crepo> --bundle \u003Creviewed-v1.json>`.\n\n## Prereq (fail-closed)\n\n`northstar` assumes the `ai-catapult-init` v3 structure is already present and never\nbootstraps it. `prereq-check.sh` asserts the `.ai\u002F` structure exists and exits\nnon-zero with guidance if absent. Do not proceed past a failed prereq. See\n[modules\u002Fprereq.md](modules\u002Fprereq.md).\n\n## The loop (delegation)\n\n`deep-interview` is the **primary** driver: interview one question at a time\nuntil ambiguity is at or below threshold. The **adversarial, skippable** pass is\ntwo skills the user declines or accepts as a unit: `grill-with-docs`\n(doc-grounded — stress-tests the plan against the repo's language, ADRs and\n`CONTEXT.md`) then `grill-me` (open-ended). \"Both satisfied\" = the\ndeep-interview gate is met AND (the adversarial pass is clear OR it was\nskipped). Delegate to those skills; do not reimplement their loops. See\n[modules\u002Floop.md](modules\u002Floop.md).\n\n## Always raise an issue\n\nAfter the loop, **always** raise an issue regardless of whether grill-me was\nskipped. Default is **local-first markdown**; a hosted tracker\n(GitHub\u002FADO\u002FGitLab\u002FJira) is used only when it is configured **and** authorized,\nfail-closed per the ai-catapult-init host-policy. Delegate to `to-issues` and\n`triage` for canonical state labels and ownership. See\n[modules\u002Fissue.md](modules\u002Fissue.md).\n\n## Ralplan → sliced goals\n\nRun `ralplan` (consensus planning) on the crystallized spec to produce **sliced\ngoals** — one tracer-bullet slice per future PR. `ralplan` owns the planning\nloop; northstar records its output as the sliced-goal artifacts.\n\n## A→B handoff\n\n`handoff-write.sh` publishes the reviewed `handoff-goals\u002F1` bundle as immutable\n`goals.json`, `handoff.md`, and `graph.json` payloads under\n`.ai\u002Fhandoff\u002Freadiness-v1\u002F\u003Cplan-id>\u002F\u003Cgeneration>\u002F`. It adds generation-addressed\nnodes and backlinks to the live `schema_version 1.1` traceability graph, then\natomically publishes `.ai\u002Fworkflows\u002Fnorthstar-readiness-v1.json` last. This\nseparate registry's `plans[]` is the sole completion pointer.\n\nThe bundle pins the spec digest and exact repository and goal identities; planning completion is **not implementation readiness**: report preparation,\nimplementation, and merge gaps separately, without claiming execution authority.\nIdempotent recovery preserves evidence; legacy migration and matched dependencies are mandatory.\nSee [modules\u002Fhandoff.md](modules\u002Fhandoff.md).\n\n## Command surface\n\n`northstar` registers as a first-class command under `.ai\u002Fcommands\u002Fomx\u002F` and `.ai\u002Fcommands\u002Fomc\u002F` using one shared schema.\nSee [modules\u002Fcommand-surface.md](modules\u002Fcommand-surface.md).\n\n## Safety rules\n\n- Planning only: never implement product changes or execute a sliced goal.\n- Planning completion is not implementation readiness; report preparation, implementation and merge gaps separately.\n- Publish completion through the separate v1 registry, never legacy discovery; missing or mismatched peers block publication.\n- Hard stop: do not invoke `autobahn`, `ultragoal`, `team`, `ralph`, `ultrawork`,\n  or another implementation engine; stop after the verified A→B handoff — never implement in the same run.\n- Fail closed: a missing prereq, missing authorization, or partial handoff write stops with guidance — never silently proceed or mutate.\n- Compose, never reimplement: delegate every loop to its owning skill.\n- Adversarial pass is `grill-with-docs` then `grill-me`, taken or declined as one\n  unit; the loop is not done on the deep-interview gate alone unless the user\n  explicitly skipped that pass.\n- Local-first: never create a hosted issue unless the tracker is configured and\n  authorized.\n\n## References\n\n- [modules\u002Fprereq.md](modules\u002Fprereq.md) — ai-catapult-init presence contract.\n- [modules\u002Floop.md](modules\u002Floop.md) — deep-interview + grill-me \"both satisfied\" rule.\n- [modules\u002Fissue.md](modules\u002Fissue.md) — local-first \u002F hosted-if-authorized issue raising.\n- [modules\u002Fhandoff.md](modules\u002Fhandoff.md) — A→B handoff schema and recovery.\n- [modules\u002Fcommand-surface.md](modules\u002Fcommand-surface.md) — shared omx\u002Fomc command schema.\n\nEnd by declaring knowledge kind `plan`, then reading\n[ai-catapult-init\u002Fmodules\u002Fknowledge-publish.md](..\u002Fai-catapult-init\u002Fmodules\u002Fknowledge-publish.md) and running its host-neutral producer step — publish when a registry exists, otherwise record `unpublished: \u003Creason>`; never fail the skill because of the registry.\n","# Northstar\n\nNorthstar converges intent into a tracked, sliced plan inside a repo initialized\nwith `ai-catapult-init` v3. This lightweight **composer** delegates to existing skills; its output is the A→B handoff that `autobahn` consumes.\n\n## Execution boundary (hard stop)\n\nNorthstar is **planning-only**. Never implement a sliced goal, modify product\ncode or tests, run implementation engines, or start `autobahn` while this skill\nis active. Writes are limited to the planning\u002Ftracking artifacts required below.\nAfter verifying the A→B handoff, **stop and report it**; implementation requires\na separate, explicit `autobahn` invocation. Never continue into implementation\nin the same run, even when the initiating prompt also asks for implementation.\n\n## Quick Start\n\n1. Run the prereq gate (fail-closed) against the repo root:\n   `bash northstar\u002Fprereq-check.sh --root .`\n2. Run the interview loop, then always raise an issue.\n3. Run `ralplan` to produce sliced goals.\n4. Publish the reviewed versioned bundle: `bash northstar\u002Fhandoff-write.sh --root \u003Crepo> --bundle \u003Creviewed-v1.json>`.\n\n## Prereq (fail-closed)\n\n`northstar` assumes the `ai-catapult-init` v3 structure is already present and never\nbootstraps it. `prereq-check.sh` asserts the `.ai\u002F` structure exists and exits\nnon-zero with guidance if absent. Do not proceed past a failed prereq. See\n[modules\u002Fprereq.md](modules\u002Fprereq.md).\n\n## The loop (delegation)\n\n`deep-interview` is the **primary** driver: interview one question at a time\nuntil ambiguity is at or below threshold. The **adversarial, skippable** pass is\ntwo skills the user declines or accepts as a unit: `grill-with-docs`\n(doc-grounded — stress-tests the plan against the repo's language, ADRs and\n`CONTEXT.md`) then `grill-me` (open-ended). \"Both satisfied\" = the\ndeep-interview gate is met AND (the adversarial pass is clear OR it was\nskipped). Delegate to those skills; do not reimplement their loops. See\n[modules\u002Floop.md](modules\u002Floop.md).\n\n## Always raise an issue\n\nAfter the loop, **always** raise an issue regardless of whether grill-me was\nskipped. Default is **local-first markdown**; a hosted tracker\n(GitHub\u002FADO\u002FGitLab\u002FJira) is used only when it is configured **and** authorized,\nfail-closed per the ai-catapult-init host-policy. Delegate to `to-issues` and\n`triage` for canonical state labels and ownership. See\n[modules\u002Fissue.md](modules\u002Fissue.md).\n\n## Ralplan → sliced goals\n\nRun `ralplan` (consensus planning) on the crystallized spec to produce **sliced\ngoals** — one tracer-bullet slice per future PR. `ralplan` owns the planning\nloop; northstar records its output as the sliced-goal artifacts.\n\n## A→B handoff\n\n`handoff-write.sh` publishes the reviewed `handoff-goals\u002F1` bundle as immutable\n`goals.json`, `handoff.md`, and `graph.json` payloads under\n`.ai\u002Fhandoff\u002Freadiness-v1\u002F\u003Cplan-id>\u002F\u003Cgeneration>\u002F`. It adds generation-addressed\nnodes and backlinks to the live `schema_version 1.1` traceability graph, then\natomically publishes `.ai\u002Fworkflows\u002Fnorthstar-readiness-v1.json` last. This\nseparate registry's `plans[]` is the sole completion pointer.\n\nThe bundle pins the spec digest and exact repository and goal identities; planning completion is **not implementation readiness**: report preparation,\nimplementation, and merge gaps separately, without claiming execution authority.\nIdempotent recovery preserves evidence; legacy migration and matched dependencies are mandatory.\nSee [modules\u002Fhandoff.md](modules\u002Fhandoff.md).\n\n## Command surface\n\n`northstar` registers as a first-class command under `.ai\u002Fcommands\u002Fomx\u002F` and `.ai\u002Fcommands\u002Fomc\u002F` using one shared schema.\nSee [modules\u002Fcommand-surface.md](modules\u002Fcommand-surface.md).\n\n## Safety rules\n\n- Planning only: never implement product changes or execute a sliced goal.\n- Planning completion is not implementation readiness; report preparation, implementation and merge gaps separately.\n- Publish completion through the separate v1 registry, never legacy discovery; missing or mismatched peers block publication.\n- Hard stop: do not invoke `autobahn`, `ultragoal`, `team`, `ralph`, `ultrawork`,\n  or another implementation engine; stop after the verified A→B handoff — never implement in the same run.\n- Fail closed: a missing prereq, missing authorization, or partial handoff write stops with guidance — never silently proceed or mutate.\n- Compose, never reimplement: delegate every loop to its owning skill.\n- Adversarial pass is `grill-with-docs` then `grill-me`, taken or declined as one\n  unit; the loop is not done on the deep-interview gate alone unless the user\n  explicitly skipped that pass.\n- Local-first: never create a hosted issue unless the tracker is configured and\n  authorized.\n\n## References\n\n- [modules\u002Fprereq.md](modules\u002Fprereq.md) — ai-catapult-init presence contract.\n- [modules\u002Floop.md](modules\u002Floop.md) — deep-interview + grill-me \"both satisfied\" rule.\n- [modules\u002Fissue.md](modules\u002Fissue.md) — local-first \u002F hosted-if-authorized issue raising.\n- [modules\u002Fhandoff.md](modules\u002Fhandoff.md) — A→B handoff schema and recovery.\n- [modules\u002Fcommand-surface.md](modules\u002Fcommand-surface.md) — shared omx\u002Fomc command schema.\n\nEnd by declaring knowledge kind `plan`, then reading\n[ai-catapult-init\u002Fmodules\u002Fknowledge-publish.md](..\u002Fai-catapult-init\u002Fmodules\u002Fknowledge-publish.md) and running its host-neutral producer step — publish when a registry exists, otherwise record `unpublished: \u003Creason>`; never fail the skill because of the registry.\n",{"eval":5,"name":5,"description":7},"main","037fe34299eb9c4a30b51f1a6bd101eae1c469b1","02-govern-plan\u002Fnorthstar","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills","https:\u002F\u002Fgithub.com\u002Fr3dlex","2026-10-29 16:22:54",{"source":8,"skillName":5,"commands":42,"prompt":52,"zipAvailable":53},[43,46,49],{"tool":44,"command":45},"npx","npx skills add r3dlex\u002Fskills --skill northstar",{"tool":47,"command":48},"bunx","bunx skills add r3dlex\u002Fskills --skill northstar",{"tool":50,"command":51},"pnpm","pnpm dlx skills add r3dlex\u002Fskills --skill northstar","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002FSKILL.md",true,[55,61,65,71,75,79,83,87,91,95,99],{"path":56,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":59,"sourceHtmlUrl":60},"approve.sh","SCRIPT","脚本",1646,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fapprove.sh",{"path":62,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":63,"sourceHtmlUrl":64},"handoff-write.sh",1506,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fhandoff-write.sh",{"path":66,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":69,"sourceHtmlUrl":70},"modules\u002Fapproval.md","REFERENCE","参考文档",4391,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fmodules\u002Fapproval.md",{"path":72,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":73,"sourceHtmlUrl":74},"modules\u002Fcommand-surface.md",3167,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fmodules\u002Fcommand-surface.md",{"path":76,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":77,"sourceHtmlUrl":78},"modules\u002Fhandoff.md",7449,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fmodules\u002Fhandoff.md",{"path":80,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":81,"sourceHtmlUrl":82},"modules\u002Fissue.md",1329,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fmodules\u002Fissue.md",{"path":84,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":85,"sourceHtmlUrl":86},"modules\u002Floop.md",2498,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fmodules\u002Floop.md",{"path":88,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":89,"sourceHtmlUrl":90},"modules\u002Fprereq.md",1287,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fmodules\u002Fprereq.md",{"path":92,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":93,"sourceHtmlUrl":94},"prereq-check.sh",1519,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Fprereq-check.sh",{"path":96,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":97,"sourceHtmlUrl":98},"readiness-dependency-v2.json",1846,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Freadiness-dependency-v2.json",{"path":100,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":101,"sourceHtmlUrl":102},"readiness-dependency.json",359,"https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002F02-govern-plan\u002Fnorthstar\u002Freadiness-dependency.json",[104,112,125,134,144,167],{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":110,"publishedAt":30},"01M6R652E42RZT7M2V3RWW991Z","skill","Fixture.","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fcatalog-malformed\u002Fskill\u002FSKILL.md",[],[111],"fixture",{"skillNo":113,"name":114,"nameZh":6,"description":115,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":116,"categories":117,"keywords":123,"publishedAt":30},"01M6R64MTPF0FGWHFS4DWP40GR","stable-skill","Fixture stable-skill.","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fcatalog-lifecycle\u002Fstable-skill\u002FSKILL.md",[118],{"categoryNo":119,"categoryName":120,"taxonomyType":121,"taxonomyTypeDesc":122},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[124,111],"stable skill",{"skillNo":126,"name":127,"nameZh":6,"description":128,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":129,"categories":130,"keywords":132,"publishedAt":30},"01M6R64HD4QVAYM2R0JSYNVDJW","experimental-skill","Fixture experimental-skill.","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fcatalog-lifecycle\u002Fexperimental-skill\u002FSKILL.md",[131],{"categoryNo":119,"categoryName":120,"taxonomyType":121,"taxonomyTypeDesc":122},[133,111],"experimental skill",{"skillNo":135,"name":136,"nameZh":6,"description":137,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":138,"categories":139,"keywords":141,"publishedAt":30},"01M6R64FZH05TEJQW75QX1E5XD","deprecated-skill","Fixture deprecated-skill.","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fcatalog-lifecycle\u002Fdeprecated-skill\u002FSKILL.md",[140],{"categoryNo":119,"categoryName":120,"taxonomyType":121,"taxonomyTypeDesc":122},[142,143,111],"deprecated skill","deprecated",{"skillNo":145,"name":146,"nameZh":6,"description":147,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":148,"categories":149,"keywords":150,"publishedAt":30},"01M6R64D13X5DC980GREWFSSFQ","karpathy-guidelines","Apply think-before-coding, simplicity, surgical-change, and goal-driven engineering rules. Use when planning, implementing, or reviewing code.","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002Freference\u002Fgolden-skills\u002F.agents\u002Fskills\u002Fkarpathy-guidelines\u002FSKILL.md",[],[151,152,153,154,25,155,156,157,158,159,160,161,162,163,164,165,20,166],"karpathy guidelines","karpathy","guidelines","Rules","simplicity","goal","think","driven","reviewing code","implementing","apply","when","coding","change","code","engineering",{"skillNo":168,"name":146,"nameZh":6,"description":147,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":169,"categories":170,"keywords":171,"publishedAt":30},"01M6R640ZS60BRV9P49FT46ZJX","https:\u002F\u002Fgithub.com\u002Fr3dlex\u002Fskills\u002Fblob\u002Fmain\u002Freference\u002Fgolden-root\u002F.agents\u002Fskills\u002Fkarpathy-guidelines\u002FSKILL.md",[],[153,151,152,164,158,165,20,166,163,154,25,155,156,157,159,160,161,162]]