[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6PVCRDXD23FA98C2DD63EN1":3,"skill-same:01M6PVCRDXD23FA98C2DD63EN1":66},{"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":32,"linkOnly":33,"markdownContent":34,"markdownBody":35,"frontMatter":36,"refName":40,"commitSha":41,"skillDirectory":42,"forksCount":39,"repositoryHtmlUrl":43,"repositoryHomepageUrl":6,"ownerHtmlUrl":44,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":45,"install":46,"resources":59},"01M6PVCRDXD23FA98C2DD63EN1","rack-advanced",null,"Advanced rack beyond day-to-day lifecycle: authoring a new custom backlog, and cross-project work. Use when defining a backlog beyond the built-in tasks\u002FHYP\u002FPROP, or registering and searching backlogs across several projects.","muratovv\u002Fai-hats","muratovv","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F7502801?v=4","01M6JGCBN4KEJX012GPJVSH04T","Fedor Muratov","01M6JGCBSNMN7CW3RGTPPE0MN2",3,"MIT","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fcore\u002Fskills\u002Frack-advanced\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31],"advanced","rack advanced","tasks","lifecycle","projects","work","use","backlog","across","project","project work","authoring","custom","2026-11-10 10:12:30",false,"---\nname: rack-advanced\ndescription: >-\n  Advanced rack beyond day-to-day lifecycle: authoring a new custom\n  backlog, and cross-project work. Use when defining a backlog beyond the\n  built-in tasks\u002FHYP\u002FPROP, or registering and searching backlogs across\n  several projects.\nlicense: MIT\nai_hats:\n  schema_version: 1\n  type: skill\n---\n\n# Rack Advanced\n\nAdvanced rack operations beyond day-to-day lifecycle:\n\n- **Authoring a new backlog** — define a `backlog.yaml` and mount it under\n  `tracker\u002F`. → [Authoring a new backlog](#authoring-a-new-backlog)\n- **Cross-project work** — sweep one backlog across projects: register them\n  ([Registry](#registry)), then list \u002F read across them\n  ([Search across projects](#search-across-projects)).\n\n## When to Use\n\nReach here when the task is to **define a new backlog** (a distinct catalog with\nits own id prefix, state machine, fields, and link kinds) — e.g. a project\n`decisions`, `experiments`, or `incidents` backlog alongside the built-in\ntasks \u002F hypotheses \u002F proposals.\n\nAlso reach here for **cross-project work** — registering projects with `rack root`\n([Registry](#registry)) or listing \u002F reading one backlog across several projects\n([Search across projects](#search-across-projects)).\n\nNot this skill when you are:\n\n- **filing a card** (`rack create`, or the `backlog-create` shim) — that adds a\n  card to an existing backlog, it does not define one;\n- **driving a card's lifecycle \u002F ops** (transition, log, link, hyp\u002Fproposal\n  verbs) — that is **hatrack**.\n\nThe moment you are editing `backlog.yaml` itself (not a `task.yaml`), you are here.\n\n## Authoring a new backlog\n\nThere is **no `rack init` \u002F register verb** — a backlog is drop-a-file. Author\nthe definition, place it under `tracker\u002F`, and it mounts on the next `rack` call.\n\n1. **Pick identity.** A `prefix` (cards become `\u003Cprefix>-\u003Cn>`) unique within the\n   root, and a `cli_alias` (the short `rack \u003Cgroup>` name) unique across mounted\n   backlogs. Do not reuse `HATS` or an existing alias (`DuplicatePrefixError` \u002F\n   `DuplicateGroupNameError` at mount).\n2. **Make the catalog dir** under `\u003Cai_hats_dir>\u002Ftracker\u002F` — convention\n   `tracker\u002F\u003Cname>\u002F`, a sibling of `backlog\u002F`. Anywhere under `tracker\u002F` works\n   except inside `backlog\u002F`, which `backlog_write_gate.py` denies to every\n   writer but `rack` (its one carve-out is `tasks\u002F\u003CID>\u002Fplan.md`).\n3. **Author `backlog.yaml`** at that dir's root. Full key grammar +\n   fail-closed invariants: **`references\u002Fbacklog-yaml-grammar.md`**. Fastest\n   start — print a shipped example (HYP \u002F PROP) from the installed package and\n   adapt (works in any project, no repo checkout needed):\n\n   ```bash\n   python -c \"from ai_hats_rack.definition import packaged_definition_source as s; print(s('hypotheses'))\"  # or: proposals\n   ```\n\n   (If a bare `python` can't import `ai_hats_rack`, use the interpreter that\n   backs `rack`.) Or start from the minimum skeleton:\n\n   ```yaml\n   name: decisions\n   prefix: DEC\n   cli_alias: decision\n   fsm:\n     initial: open\n     states: [{ name: open }, { name: accepted }, { name: superseded }]\n     edges:\n       - { from: open, to: accepted, name: accept }\n       - { from: open, to: superseded, name: supersede }\n   links:\n     kinds: # must be NON-EMPTY; a symmetric self-inverse kind is the minimum\n       - { name: related, arity: many, inverse: related }\n   fields:\n     - { name: context, type: str, default: \"\" }\n   ```\n\n4. **It auto-mounts.** `Workspace.discover` → `_scan_sibling_backlogs` walks\n   `tracker\u002F**` and picks up every `backlog.yaml`. No registration step.\n5. **Verify it landed:** `rack --help` lists the `\u003Ccli_alias>` group; `rack\n   \u003Calias> create \"…\"` mints `\u003CPREFIX>-001` (ids are zero-padded, min 3\n   digits); `rack context \u003CPREFIX>-001` routes by prefix; `rack ls --backlog\n   \u003Calias>` filters to it.\n\n**Before you save — invariant checklist** (each is a fail-closed loader error):\n\n- Every top\u002Fsection key is in the grammar — a stray key (e.g. `sections:`) is\n  rejected (`UnsupportedBacklogKeyError`).\n- `links.kinds` is **non-empty** (`LinksRegistryError` otherwise) — a backlog\n  with no cross-links still needs one kind; `related` (symmetric) is the minimum.\n- No edge `name` equals a state name (`EdgeNameStateCollisionError`).\n- Any **stored inverse pair** (`inverse:` on both, both stored) carries\n  `handlers: [mirror-link]` (`MissingMirrorReactionError`). Symmetric\n  (`related`) and derived (`children`) kinds are exempt.\n- Complex fields are `type: any` **plus** a `validator:` name.\n- Every handler \u002F validator \u002F extension name resolves in the stock code-side\n  registry — you cannot introduce one in YAML alone.\n\n**Plan sections are separate.** If (and only if) the backlog uses the plan gate,\nits plan-section catalog lives in a sibling `plan-sections.yaml`, **not** a\n`backlog.yaml` key (never-drift). Most custom backlogs need neither.\n\n## What the backlog gets for free\n\nOnce mounted, the definition alone yields (no per-backlog code):\n\n- `rack \u003Calias> create …` — schema-driven from `fields` (required\u002Fchoices enforced).\n- `rack \u003Calias> update \u003CID> --\u003Cfield> …` — scalar (str\u002Fint) field edits.\n- Extension verbs any declared `extensions` contribute (e.g. `append-verdict`).\n- Prefix-routed reads\u002Fmoves: `rack context \u003CID>`, `rack ls \u003CID>`, `rack\n  transition \u003CID> \u003Cstate|edge-name>`, `rack ls --backlog \u003Calias>`.\n\n## Registry\n\nTo sweep a backlog across projects (below), the projects must be known. Note the\ncontrast with [Authoring a new backlog](#authoring-a-new-backlog): a *backlog*\nneeds no registration (drop a `backlog.yaml` and it auto-mounts), but a *project*\nis registered explicitly — a persistent list at `~\u002F.ai-hats\u002Froots.yaml` (override\nthe path with `RACK_ROOTS_FILE`):\n\n```bash\nrack root add \u003Cpath>   # register a project root (must hold .agent\u002F or ai-hats.yaml)\nrack root ls           # list registered roots: root_id, path, reachable?\nrack root rm \u003Cpath>    # unregister a root by path\n```\n\n`root_id` is the project's **folder name** (no alias). `add` validates the path is\na project, stores it resolved, and de-dupes (a re-add is a no-op); `rm` takes the\nsame path. Registering is optional — `--root` (below) sweeps ad-hoc roots without\ntouching the file.\n\n## Search across projects\n\nList or read **one backlog across several projects** — \"hypotheses of all\nprojects\" in one command — over the registered roots ([Registry](#registry))\nand\u002For ad-hoc `--root` paths.\n\n**List** on the no-id scan (rows gain a `project` marker column \u002F `project` json\nkey; the current project is **always** included, roots dedup by real path):\n\n```bash\nrack ls --projects all               # every registered project ∪ the current one\nrack ls --projects projA,projB       # a named subset (by root_id) ∪ the current one\nrack ls --root ..\u002Fother              # ad-hoc: add a root by path (repeatable)\nrack ls --backlog hyp --projects all # the hyp backlog in EVERY swept project\n```\n\n**Read** — a bare id is ambiguous when two projects share a prefix, so qualify it\n(or mount an unregistered project by path):\n\n```bash\nrack context projB:ACME-9              # \u003Croot_id>:\u003Cid> routes into a registered project\nrack context projB:ACME-9 --root ..\u002FprojB   # or mount it by path\n```\n\n**Boundaries.**\n\n- **Read-only.** The sweep and qualified `context` are reads; cross-project\n  *writes* (qualified `transition`, cross-project `create`, cross-project links)\n  are out of scope — a sweep never mutates another project.\n- **Read-tolerant registry, fail-fast `--root`.** A registered root that has\n  vanished is *skipped* with a non-silent footer (the sweep survives); an explicit\n  `--root` to a non-project is a hard error.\n- **Folder-name collision** (two `ai-hats` clones): the `\u003Croot_id>:\u003Cid>` qualifier\n  is ambiguous — disambiguate with the full path via `--root`.\n\n## Completion\n\n**Authoring a backlog** — done when: the `backlog.yaml` sits under `tracker\u002F`,\n`rack --help` shows its group, and `rack \u003Calias> create` + `rack context\n\u003CPREFIX>-001` both work — with no loader error on any `rack` call.\n\n**Cross-project sweep** — done when: `rack root add \u003Cpath>` then `rack root ls`\nshows it, and `rack ls --projects all` lists cards from the registered project(s)\nalongside the current one, each row carrying its `project` marker.\n\n**Validation scenario (RED → GREEN).** RED — asked to add a project `decisions`\nbacklog, an agent *without* this skill hunts for a `rack init-backlog`\u002Fregister\nverb that does not exist, reuses the `HATS` prefix (`DuplicatePrefixError`), puts\na `sections:` key in the file (rejected — unknown key), or declares a\n`supersedes`\u002F`superseded_by` pair with no `mirror-link`\n(`MissingMirrorReactionError`) — a multi-error discovery loop. GREEN — with the\nskill it writes `tracker\u002Fdecisions\u002Fbacklog.yaml` (unique `prefix: DEC` +\n`cli_alias: decision`, valid fsm\u002Ffields\u002Flinks, no `sections:`, `mirror-link` on\nany stored inverse), drops it in, and `rack decision create \"…\"` works first try.\n\n## Anti-Patterns\n\n- Searching for a create\u002Fregister\u002Finit command — mounting is drop-a-file;\n  `Workspace.discover` does it.\n- Reusing `HATS` (or another backlog's prefix \u002F alias) — collisions fail closed\n  at mount.\n- Putting a `sections:` (or any undeclared) key in `backlog.yaml` — plan sections\n  are a sibling `plan-sections.yaml`; unknown keys are rejected.\n- Declaring a stored inverse link pair without `mirror-link` — the reverse edge\n  would drift, so the loader refuses it.\n- Inventing a handler \u002F validator name in YAML with no code-side factory — the\n  name must resolve in the stock registry or composition fails closed.\n- Expecting `--projects` \u002F `--root` to route a *write* — the sweep is read-only;\n  cross-project mutation is out of scope.\n- Leaning on a `\u003Croot_id>:\u003Cid>` qualifier when two roots share a folder name — the\n  id is ambiguous; use the full path via `--root`.\n- Hand-editing `~\u002F.ai-hats\u002Froots.yaml` for a bad path — `rack root add` validates\n  and resolves; a non-project path is rejected there.\n","# Rack Advanced\n\nAdvanced rack operations beyond day-to-day lifecycle:\n\n- **Authoring a new backlog** — define a `backlog.yaml` and mount it under\n  `tracker\u002F`. → [Authoring a new backlog](#authoring-a-new-backlog)\n- **Cross-project work** — sweep one backlog across projects: register them\n  ([Registry](#registry)), then list \u002F read across them\n  ([Search across projects](#search-across-projects)).\n\n## When to Use\n\nReach here when the task is to **define a new backlog** (a distinct catalog with\nits own id prefix, state machine, fields, and link kinds) — e.g. a project\n`decisions`, `experiments`, or `incidents` backlog alongside the built-in\ntasks \u002F hypotheses \u002F proposals.\n\nAlso reach here for **cross-project work** — registering projects with `rack root`\n([Registry](#registry)) or listing \u002F reading one backlog across several projects\n([Search across projects](#search-across-projects)).\n\nNot this skill when you are:\n\n- **filing a card** (`rack create`, or the `backlog-create` shim) — that adds a\n  card to an existing backlog, it does not define one;\n- **driving a card's lifecycle \u002F ops** (transition, log, link, hyp\u002Fproposal\n  verbs) — that is **hatrack**.\n\nThe moment you are editing `backlog.yaml` itself (not a `task.yaml`), you are here.\n\n## Authoring a new backlog\n\nThere is **no `rack init` \u002F register verb** — a backlog is drop-a-file. Author\nthe definition, place it under `tracker\u002F`, and it mounts on the next `rack` call.\n\n1. **Pick identity.** A `prefix` (cards become `\u003Cprefix>-\u003Cn>`) unique within the\n   root, and a `cli_alias` (the short `rack \u003Cgroup>` name) unique across mounted\n   backlogs. Do not reuse `HATS` or an existing alias (`DuplicatePrefixError` \u002F\n   `DuplicateGroupNameError` at mount).\n2. **Make the catalog dir** under `\u003Cai_hats_dir>\u002Ftracker\u002F` — convention\n   `tracker\u002F\u003Cname>\u002F`, a sibling of `backlog\u002F`. Anywhere under `tracker\u002F` works\n   except inside `backlog\u002F`, which `backlog_write_gate.py` denies to every\n   writer but `rack` (its one carve-out is `tasks\u002F\u003CID>\u002Fplan.md`).\n3. **Author `backlog.yaml`** at that dir's root. Full key grammar +\n   fail-closed invariants: **`references\u002Fbacklog-yaml-grammar.md`**. Fastest\n   start — print a shipped example (HYP \u002F PROP) from the installed package and\n   adapt (works in any project, no repo checkout needed):\n\n   ```bash\n   python -c \"from ai_hats_rack.definition import packaged_definition_source as s; print(s('hypotheses'))\"  # or: proposals\n   ```\n\n   (If a bare `python` can't import `ai_hats_rack`, use the interpreter that\n   backs `rack`.) Or start from the minimum skeleton:\n\n   ```yaml\n   name: decisions\n   prefix: DEC\n   cli_alias: decision\n   fsm:\n     initial: open\n     states: [{ name: open }, { name: accepted }, { name: superseded }]\n     edges:\n       - { from: open, to: accepted, name: accept }\n       - { from: open, to: superseded, name: supersede }\n   links:\n     kinds: # must be NON-EMPTY; a symmetric self-inverse kind is the minimum\n       - { name: related, arity: many, inverse: related }\n   fields:\n     - { name: context, type: str, default: \"\" }\n   ```\n\n4. **It auto-mounts.** `Workspace.discover` → `_scan_sibling_backlogs` walks\n   `tracker\u002F**` and picks up every `backlog.yaml`. No registration step.\n5. **Verify it landed:** `rack --help` lists the `\u003Ccli_alias>` group; `rack\n   \u003Calias> create \"…\"` mints `\u003CPREFIX>-001` (ids are zero-padded, min 3\n   digits); `rack context \u003CPREFIX>-001` routes by prefix; `rack ls --backlog\n   \u003Calias>` filters to it.\n\n**Before you save — invariant checklist** (each is a fail-closed loader error):\n\n- Every top\u002Fsection key is in the grammar — a stray key (e.g. `sections:`) is\n  rejected (`UnsupportedBacklogKeyError`).\n- `links.kinds` is **non-empty** (`LinksRegistryError` otherwise) — a backlog\n  with no cross-links still needs one kind; `related` (symmetric) is the minimum.\n- No edge `name` equals a state name (`EdgeNameStateCollisionError`).\n- Any **stored inverse pair** (`inverse:` on both, both stored) carries\n  `handlers: [mirror-link]` (`MissingMirrorReactionError`). Symmetric\n  (`related`) and derived (`children`) kinds are exempt.\n- Complex fields are `type: any` **plus** a `validator:` name.\n- Every handler \u002F validator \u002F extension name resolves in the stock code-side\n  registry — you cannot introduce one in YAML alone.\n\n**Plan sections are separate.** If (and only if) the backlog uses the plan gate,\nits plan-section catalog lives in a sibling `plan-sections.yaml`, **not** a\n`backlog.yaml` key (never-drift). Most custom backlogs need neither.\n\n## What the backlog gets for free\n\nOnce mounted, the definition alone yields (no per-backlog code):\n\n- `rack \u003Calias> create …` — schema-driven from `fields` (required\u002Fchoices enforced).\n- `rack \u003Calias> update \u003CID> --\u003Cfield> …` — scalar (str\u002Fint) field edits.\n- Extension verbs any declared `extensions` contribute (e.g. `append-verdict`).\n- Prefix-routed reads\u002Fmoves: `rack context \u003CID>`, `rack ls \u003CID>`, `rack\n  transition \u003CID> \u003Cstate|edge-name>`, `rack ls --backlog \u003Calias>`.\n\n## Registry\n\nTo sweep a backlog across projects (below), the projects must be known. Note the\ncontrast with [Authoring a new backlog](#authoring-a-new-backlog): a *backlog*\nneeds no registration (drop a `backlog.yaml` and it auto-mounts), but a *project*\nis registered explicitly — a persistent list at `~\u002F.ai-hats\u002Froots.yaml` (override\nthe path with `RACK_ROOTS_FILE`):\n\n```bash\nrack root add \u003Cpath>   # register a project root (must hold .agent\u002F or ai-hats.yaml)\nrack root ls           # list registered roots: root_id, path, reachable?\nrack root rm \u003Cpath>    # unregister a root by path\n```\n\n`root_id` is the project's **folder name** (no alias). `add` validates the path is\na project, stores it resolved, and de-dupes (a re-add is a no-op); `rm` takes the\nsame path. Registering is optional — `--root` (below) sweeps ad-hoc roots without\ntouching the file.\n\n## Search across projects\n\nList or read **one backlog across several projects** — \"hypotheses of all\nprojects\" in one command — over the registered roots ([Registry](#registry))\nand\u002For ad-hoc `--root` paths.\n\n**List** on the no-id scan (rows gain a `project` marker column \u002F `project` json\nkey; the current project is **always** included, roots dedup by real path):\n\n```bash\nrack ls --projects all               # every registered project ∪ the current one\nrack ls --projects projA,projB       # a named subset (by root_id) ∪ the current one\nrack ls --root ..\u002Fother              # ad-hoc: add a root by path (repeatable)\nrack ls --backlog hyp --projects all # the hyp backlog in EVERY swept project\n```\n\n**Read** — a bare id is ambiguous when two projects share a prefix, so qualify it\n(or mount an unregistered project by path):\n\n```bash\nrack context projB:ACME-9              # \u003Croot_id>:\u003Cid> routes into a registered project\nrack context projB:ACME-9 --root ..\u002FprojB   # or mount it by path\n```\n\n**Boundaries.**\n\n- **Read-only.** The sweep and qualified `context` are reads; cross-project\n  *writes* (qualified `transition`, cross-project `create`, cross-project links)\n  are out of scope — a sweep never mutates another project.\n- **Read-tolerant registry, fail-fast `--root`.** A registered root that has\n  vanished is *skipped* with a non-silent footer (the sweep survives); an explicit\n  `--root` to a non-project is a hard error.\n- **Folder-name collision** (two `ai-hats` clones): the `\u003Croot_id>:\u003Cid>` qualifier\n  is ambiguous — disambiguate with the full path via `--root`.\n\n## Completion\n\n**Authoring a backlog** — done when: the `backlog.yaml` sits under `tracker\u002F`,\n`rack --help` shows its group, and `rack \u003Calias> create` + `rack context\n\u003CPREFIX>-001` both work — with no loader error on any `rack` call.\n\n**Cross-project sweep** — done when: `rack root add \u003Cpath>` then `rack root ls`\nshows it, and `rack ls --projects all` lists cards from the registered project(s)\nalongside the current one, each row carrying its `project` marker.\n\n**Validation scenario (RED → GREEN).** RED — asked to add a project `decisions`\nbacklog, an agent *without* this skill hunts for a `rack init-backlog`\u002Fregister\nverb that does not exist, reuses the `HATS` prefix (`DuplicatePrefixError`), puts\na `sections:` key in the file (rejected — unknown key), or declares a\n`supersedes`\u002F`superseded_by` pair with no `mirror-link`\n(`MissingMirrorReactionError`) — a multi-error discovery loop. GREEN — with the\nskill it writes `tracker\u002Fdecisions\u002Fbacklog.yaml` (unique `prefix: DEC` +\n`cli_alias: decision`, valid fsm\u002Ffields\u002Flinks, no `sections:`, `mirror-link` on\nany stored inverse), drops it in, and `rack decision create \"…\"` works first try.\n\n## Anti-Patterns\n\n- Searching for a create\u002Fregister\u002Finit command — mounting is drop-a-file;\n  `Workspace.discover` does it.\n- Reusing `HATS` (or another backlog's prefix \u002F alias) — collisions fail closed\n  at mount.\n- Putting a `sections:` (or any undeclared) key in `backlog.yaml` — plan sections\n  are a sibling `plan-sections.yaml`; unknown keys are rejected.\n- Declaring a stored inverse link pair without `mirror-link` — the reverse edge\n  would drift, so the loader refuses it.\n- Inventing a handler \u002F validator name in YAML with no code-side factory — the\n  name must resolve in the stock registry or composition fails closed.\n- Expecting `--projects` \u002F `--root` to route a *write* — the sweep is read-only;\n  cross-project mutation is out of scope.\n- Leaning on a `\u003Croot_id>:\u003Cid>` qualifier when two roots share a folder name — the\n  id is ambiguous; use the full path via `--root`.\n- Hand-editing `~\u002F.ai-hats\u002Froots.yaml` for a bad path — `rack root add` validates\n  and resolves; a non-project path is rejected there.\n",{"name":5,"ai_hats":37,"license":15,"description":7},{"type":38,"schema_version":39},"skill",1,"master","f922a7f627276366873d546942d9ad899a86f71a","packages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fcore\u002Fskills\u002Frack-advanced","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats","https:\u002F\u002Fgithub.com\u002Fmuratovv","2026-11-05 20:59:10",{"source":8,"skillName":5,"commands":47,"prompt":57,"zipAvailable":58},[48,51,54],{"tool":49,"command":50},"npx","npx skills add muratovv\u002Fai-hats --skill rack-advanced",{"tool":52,"command":53},"bunx","bunx skills add muratovv\u002Fai-hats --skill rack-advanced",{"tool":55,"command":56},"pnpm","pnpm dlx skills add muratovv\u002Fai-hats --skill rack-advanced","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fcore\u002Fskills\u002Frack-advanced\u002FSKILL.md",true,[60],{"path":61,"resourceType":62,"resourceTypeDesc":63,"sizeBytes":64,"sourceHtmlUrl":65},"references\u002Fbacklog-yaml-grammar.md","REFERENCE","参考文档",11774,"https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fcore\u002Fskills\u002Frack-advanced\u002Freferences\u002Fbacklog-yaml-grammar.md",[67,92,117,140,164,178],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":32},"01M6PVFB1SM738R2SRSRSF7CY6","tool-evaluation-protocol","Time-bounded protocol for evaluating new tools, libraries, or methodologies before adoption. Use when evaluating a candidate dependency for the project, researching a methodology (e.g. ML\u002FLLM frameworks, testing approaches), or following up on a hype-driven recommendation.","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fusage\u002Fskills\u002Ftool-evaluation-protocol\u002FSKILL.md",[],[74,75,76,77,78,79,80,81,82,83,25,84,85,86,87,28,88,89,90,91],"protocol","Evaluation","tool","tool evaluation protocol","testing","ml","libraries","adoption","methodology","Tools","dependency","up","frameworks","time","recommendation","for","bounded","candidate",{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":32},"01M6PVF7HATBEX1G7N50E2QT57","systematic-debugging","4-phase bug-fix protocol (evidence, pattern, hypothesis, verify). Use when investigating any error, failure, or unexpected behavior, handling bug reports from users or monitoring, or diagnosing test failures with non-obvious causes.","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fusage\u002Fskills\u002Fsystematic-debugging\u002FSKILL.md",[],[99,100,101,102,74,103,104,105,106,107,25,108,109,110,111,112,113,114,115,116],"systematic debugging","debugging","systematic","test","monitoring","hypothesis","verify","behavior","fix","evidence","bug","reports","pattern","users","error","with","bug-fix","phase",{"skillNo":118,"name":119,"nameZh":6,"description":120,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":121,"categories":122,"keywords":128,"publishedAt":32},"01M6PVF3MTQN3RPF73AN34856J","skill-template","Canonical template and validation guide for ai-hats skills. Use when creating a new skill, reviewing an existing one for structural compliance, or deciding which pattern fits a new behavior.","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fusage\u002Fskills\u002Fskill-template\u002FSKILL.md",[123],{"categoryNo":124,"categoryName":125,"taxonomyType":126,"taxonomyTypeDesc":127},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[129,130,131,132,133,106,134,25,135,111,136,137,138,89,139],"template","skill template","validation","compliance","hats","guide","new skill","structural","canonical","which","Existing",{"skillNo":141,"name":142,"nameZh":6,"description":143,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":144,"categories":145,"keywords":147,"publishedAt":32},"01M6PVF06SCV28QARM5MQPANC9","skill-optimization","Audit and refactor library components to eliminate redundancy and staleness. Use during periodic library maintenance, after adding multiple new skills or rules, when a role's token budget is too high, or after retrospectives identify component issues.","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fusage\u002Fskills\u002Fskill-optimization\u002FSKILL.md",[146],{"categoryNo":124,"categoryName":125,"taxonomyType":126,"taxonomyTypeDesc":127},[148,149,150,151,152,153,154,155,156,157,25,158,159,160,161,162,163],"Optimization","skill optimization","components","component","issues","library","audit","refactor","Budget","Rules","periodic","maintenance","token","token budget","role","high",{"skillNo":165,"name":166,"nameZh":6,"description":167,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":168,"categories":169,"keywords":170,"publishedAt":32},"01M6PVEXM9717016T9MW82HDCT","silent-fallback-gate","PreToolUse hook — refuses an Edit or Write that adds an inert broad except. Use when a Python edit was blocked for a silent fallback.","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fusage\u002Fskills\u002Fsilent-fallback-gate\u002FSKILL.md",[],[171,172,173,174,175,176,25,177,89],"gate","fallback","silent fallback gate","python","write","edit","hook",{"skillNo":179,"name":180,"nameZh":6,"description":181,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":182,"categories":183,"keywords":185,"publishedAt":32},"01M6PVESM49C5NT7SB7N3F76VS","retro-to-framework","Convert an observation about agent behavior — from a retrospective or from a single incident — into a framework-level change (hook, rule, skill, skill update). Use after a retrospective identifies problems that are not project-specific, after an incident where an agent misapplied a component it had read, when CLAUDE.md band-aids accumulate (more than 3 per-project rules that could be generic), or when the same problem recurs across multiple projects.","https:\u002F\u002Fgithub.com\u002Fmuratovv\u002Fai-hats\u002Fblob\u002Fmaster\u002Fpackages\u002Fai-hats-library\u002Fsrc\u002Fai_hats_library\u002Fexperimental\u002Fusage\u002Fskills\u002Fretro-to-framework\u002FSKILL.md",[184],{"categoryNo":124,"categoryName":125,"taxonomyType":126,"taxonomyTypeDesc":127},[186,187,188,189,151,190,23,191,157,106,25,192,27,28,193,194,195,196,197,198],"framework","retro","retro to framework","change","retrospective","convert","incident","problem","about","more","skill update","single","generic"]