[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77MCVRGCDKCJW6VKWA2YYQ3":3,"skill-same:01M77MCVRGCDKCJW6VKWA2YYQ3":82},{"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":32,"commitSha":33,"skillDirectory":34,"forksCount":35,"repositoryHtmlUrl":36,"repositoryHomepageUrl":6,"ownerHtmlUrl":37,"ownerBlogUrl":38,"ownerDescription":39,"updatedAt":40,"install":41,"resources":53},"01M77MCVRGCDKCJW6VKWA2YYQ3","audit",null,"Coordinate Git\u002FGitHub hygiene across a cross-repository fleet: discover canonical repositories, collect and roll up cross-repository evidence (including merged remote-tracking heads still on origin), and hand an action plan to repo-hygiene\u002Fsource-control, which own per-repository cleanup. The current collector is read-only and emits detailed exact handoffs; it never deletes, prunes, repairs, fetches, checks out, or rewrites. Use when: 'audit repositories across a fleet', 'stale branches across repos', 'orphaned worktrees across repos', 'merged remote branches still on origin', 'moved or renamed GitHub repos'.","melodic-software\u002Fclaude-code-plugins","melodic-software","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F58273638?v=4","01M77JVQTXNGKCE4MDFY2BC936","Melodic Software","01M77JVQVN9QDWKAH6W03ANP76",30,"MIT","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\ndescription: \"Coordinate Git\u002FGitHub hygiene across a cross-repository fleet: discover canonical repositories, collect and roll up cross-repository evidence (including merged remote-tracking heads still on origin), and hand an action plan to repo-hygiene\u002Fsource-control, which own per-repository cleanup. The current collector is read-only and emits detailed exact handoffs; it never deletes, prunes, repairs, fetches, checks out, or rewrites. Use when: 'audit repositories across a fleet', 'stale branches across repos', 'orphaned worktrees across repos', 'merged remote branches still on origin', 'moved or renamed GitHub repos'.\"\nuser-invocable: true\ndisable-model-invocation: false\nargument-hint: \"[[\u003Cdir>]... [--root \u003Cdir>]... [--repo \u003Cdir>]... [--config \u003Cfile>] [--detail]|--apply-plan \u003Cpath>]\"\nallowed-tools:\n  - Bash(${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-fleet.sh:*)\nmetadata:\n  workflow-stage: operator\n  summary: Discover a repository fleet and coordinate read-only evidence handoffs\n  cadence: weekly\n---\n\n**Arguments.** `[[\u003Cdir>]... [--root \u003Cdir>]... [--repo \u003Cdir>]... [--config \u003Cfile>] [--detail]|--apply-plan \u003Cpath>]`. Full form: [[\u003Cdir>]... [--root \u003Cdir>]... [--repo \u003Cdir>]... [--config \u003Cfile>] [--canonical \u003Cgithub.com\u002Fowner\u002Frepo=path>]... [--skip \u003Cname>]... [--extend-skip \u003Cname>]... [--max-depth \u003C1..12>] [--detail] [--plan-file \u003Cpath>]|--apply-plan \u003Cpath>]\n\n## Purpose\n\nCoordinate cross-repository hygiene. This skill owns bounded fleet discovery, canonical\ncheckout resolution, fleet-scale evidence collection, rollup, and action-plan routing. It does\n**not** own per-repository cleanup decisions or execution; those belong to `repo-hygiene` and\n`source-control`.\n\nThe collector produces the detailed read-only report described below, including the compact\nmachine-readable rollup and action-plan artifact. Tell the user to run\n`\u002Frepo-fleet-hygiene:apply --plan-file \u003Cpath>` (dry-run by default; `--apply` plus confirmation or\n`--yes` to mutate; `--remote-branches` additionally opts in to per-branch remote deletion, which\n`--yes` never answers). Do not add an execute flag to this audit script.\n\n## Non-negotiable boundary\n\nNever run or suggest running inline from this skill: `git fetch`, `git worktree prune`,\n`git worktree repair`, `git worktree remove`, `git branch -d\u002F-D`, `git remote set-url`, or any\nfilesystem deletion. The bundled script has no mutation mode. `--apply-plan` is a read-only\ndry-run approval artifact over a prior plan file. A report may name a command\u002Ftool as a future\nhandoff; the fleet action plan lists those invocations once per repository behind one confirmation\ngate. Actual fleet mutation belongs to `\u002Frepo-fleet-hygiene:apply`, not this skill.\n\n## Input resolution\n\nParse `$ARGUMENTS` as opaque arguments for the bundled script. Supported flags:\n\n- `\u003Cdir>`: a bare positional path, treated as `--root`. A drive root (`D:`, `D:\u002F`) is a legitimate\n  discovery root and normalizes to `D:\u002F`. This is the form `\u002Frepo-fleet-hygiene:audit D:` uses.\n- `--root \u003Cdir>`: bounded recursive repository discovery (repeatable).\n- `--repo \u003Cdir>`: exact repository\u002Fworktree target (repeatable).\n- `--config \u003Cfile>`: explicit Git-format config (at most one).\n- `--canonical \u003Cgithub.com\u002Fowner\u002Frepo=path>`: invocation-specific canonical checkout override\n  (repeatable; explicit wins over config).\n- `--skip \u003Cname>`: discovery directory-name skip (repeatable). Explicit `--skip` \u002F `fleet.skip`\n  entries **replace** the default skip list rather than appending. Otherwise shrinking is\n  impossible. Default (neither CLI nor config): `vendor` plus the package-manager cache trees\n  `node_modules`, `.venv`, `.pnpm-store`, `.yarn`, `.npm`, `.cargo`, `.rustup`, `.gradle`, `.m2`,\n  `.nuget`, `__pycache__`, and `.tox`. To shrink (e.g. reach a repo under `vendor\u002F`), list only the\n  names you still want skipped. CLI and config compose additively with each other like other scope\n  inputs. Values must be bare directory names (no empty value, no path separator). `.`, `..`, and\n  `.git` stay skipped unconditionally even when an explicit list omits them.\n- `--extend-skip \u003Cname>`: discovery directory-name skip that **adds** to whichever list is in\n  effect (the defaults, or an explicit `--skip` \u002F `fleet.skip` list) instead of replacing it\n  (repeatable). Config equivalent: repeatable `fleet.skipAppend`. Same bare-name validation.\n- `--max-depth \u003C1..12>`: discovery bound; explicit wins over config\u002Fdefault `5`.\n- `--project-dir \u003Cdir>`: the session's project directory, used for the project-scoped config rung.\n  It is **not** a scope fallback; see the no-scope ladder below.\n- `--detail`: emit collapsed per-target evidence after the rollup (default is rollup + action plan\n  only).\n- `--plan-file \u003Cpath>`: write the machine-readable action-plan JSON to this path (otherwise a temp\n  file is created and named in the report).\n- `--apply-plan \u003Cpath>`: standalone read-only mode. Render the ordered dry-run approval artifact\n  for a previously written plan (cannot combine with discovery flags).\n\nAlways pass `--project-dir \"${CLAUDE_PROJECT_DIR}\"` on audit runs (not on `--apply-plan`). The\nscript does not read the project directory from its environment, so passing it in is what makes\nthe project rung below reachable at all.\n\n- **Pointer**: where the token is substituted is in\n  [Available string substitutions](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fskills#available-string-substitutions),\n  and which processes receive the variable is in\n  [Where each variable resolves](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmanifest-reference#where-each-variable-resolves).\n\nIf no explicit scope and no config-supplied `fleet.root`\u002F`fleet.repo` resolve, the run uses the\nshared no-scope ladder: `--named` paths, then `ghq root --all` when `ghq` is installed, then the\nnearest of the 4 parents above the working directory's checkout (above the working directory itself\noutside a checkout) that directly holds 2 or more Git repositories, so a run inside one checkout\ncovers its sibling fleet, then the checkout holding the working directory, else exit 3 naming every\nrung. A Git checkout in the working directory is a rung; the session project directory is not one\non its own. **Claim:** `ghq root --all` prints every configured root. **Basis:** the\n[ghq README](https:\u002F\u002Fgithub.com\u002Fx-motemen\u002Fghq#commands) (`ghq root [--all]`; \"Without '--all' option,\nthe primary one is shown\") and `ghq help root` on ghq 1.10.1 (\"--all  Show all roots\"). **As of:**\n2026-11-10. **Recheck:** the README's `root` entry drops or renames `--all`. Pass that guidance\nthrough rather than re-deriving a root yourself. Config resolution is the script's own ladder. Do not\npre-resolve or pass a probed path yourself:\nexplicit `--config` wins, else the script probes\n`\u003Cproject-dir>\u002F.claude\u002Frepo-fleet-hygiene.conf` (project-scoped), else\n`~\u002F.claude\u002Frepo-fleet-hygiene.conf` (user-global, a machine-scoped fleet config placed there is\nrecorded user intent, not a guessed root). The report header names the consumed config and its\nsource, or states that none was consumed. Never guess a broader machine root from the current\npath beyond that ladder.\n\nConfig-supplied scope is **additive** to CLI-supplied scope: a `--repo X` run still walks every\nconfigured root. The header's `Scope:` line names each contributing rung and its entry count, so\nreport that line rather than assuming the arguments were the whole scope.\n\nBefore execution, reject any arguments outside this grammar, noting that a bare positional path\n**is** in the grammar, so `\u002Frepo-fleet-hygiene:audit D:` and\n`\u002Frepo-fleet-hygiene:audit \u002Fpath\u002Fto\u002Ftree` are valid invocations to pass through, not arguments to\nrefuse. What stays rejected is an unrecognized flag: anything beginning with `-` that is not listed\nabove. Pass every path\u002Foverride as a quoted argument; never assemble a shell fragment from config,\nrepository, remote, or branch text.\n\nRun exactly once:\n\n```bash\n${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-fleet.sh --project-dir \"${CLAUDE_PROJECT_DIR}\" \u003Cvalidated-and-quoted-arguments>\n```\n\nThe script validates config with `git config --file`; it never sources or executes it.\n\n## Evidence rules\n\nThe bundled collector is authoritative for classifications. Preserve its evidence in the report:\n\n1. **Canonical checkout:** explicit remote-keyed override → configured remote-keyed override → the\n   repository's **main worktree**, read from `git worktree list --porcelain`, never from the\n   checkout discovery reached first. Which record is the main worktree: git-worktree's\n   [`list`](https:\u002F\u002Fgit-scm.com\u002Fdocs\u002Fgit-worktree#Documentation\u002Fgit-worktree.txt-list) entry and\n   [List output format](https:\u002F\u002Fgit-scm.com\u002Fdocs\u002Fgit-worktree#_list_output_format). When a supplied or discovered path resolves to a different main\n   worktree, the header states the substitution on one `Resolved to main worktree:` line per\n   repository, naming every path that resolved into it. Relay it, because the operator named one\n   path and the report is about another. The report always shows\n   discovered and canonical paths. An\n   override target with a missing\u002Fnon-GitHub remote, or a different identity that cannot be proven to\n   resolve to the same GitHub repository, stops that repository's local audit before evidence combines.\n2. **GitHub identity:** read the selected fetch remote with `git remote get-url`; accept only\n   `github.com\u002Fowner\u002Frepo`; query `GET \u002Frepos\u002F{owner}\u002F{repo}`. If returned `full_name` differs, report\n   `HIGH` transfer\u002Frename evidence and **continue** branch\u002Fworktree analysis against that resolved\n   identity. A moved remote is not a reason to skip local classification or merge evidence. A\n   404\u002F403\u002Fnetwork error is `UNKNOWN`, never \"deleted\" or \"moved\". A 404\u002F403 on an identity listed\n   in `fleet.ackUnavailable` is demoted to `ACKNOWLEDGED`, still reported, never suppressed; acks\n   never touch non-404\u002F403 failures or successful-response evidence.\n3. **Merged branch:** one aliased `gh api graphql` query per repository page of local branches\n   (up to `MERGED_PR_GRAPHQL_ALIAS_PAGE` `headRefName` aliases per call, `first:1`,\n   `states:[MERGED]`). GraphQL's `headRefName` argument is an **exact** match, never the search\n   API's prefix-matching `head:` qualifier, so `feature\u002Fauth` and `feature\u002Fauth-v2` never conflate.\n   Measured rate cost stays 1 per call (nodeCount equals the alias count), well under GitHub's\n   limits:\n   [Primary rate limit](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fgraphql\u002Foverview\u002Frate-limits-and-query-limits-for-the-graphql-api#primary-rate-limit)\n   and [Node limit](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fgraphql\u002Foverview\u002Frate-limits-and-query-limits-for-the-graphql-api#node-limit),\n   as of 2026-11-03; recheck when either section changes its limit. There is no merged-PR\n   result window and no per-branch fallback: every non-default local branch the operator asked\n   about is queried by exact name, including heads GitHub auto-deleted and a later fetch pruned. Fail closed when `gh`\u002FGraphQL is\n   unavailable. Emit `github-pr-evidence-unavailable` and never infer unmerged from a missing\n   row after a failed page. Identical branch names in another repository are unrelated. `HIGH`\n   requires the PR `headRefOid` to equal the current local tip. Tip drift is `MEDIUM` manual\n   review. Git ancestry without GitHub evidence is `LOW` and never called merged-by-PR, and\n   under squash merges that ancestry predicate is near-inert, so on a squash-merging fleet\n   GitHub evidence is effectively the only merge evidence.\n4. **Merged remote branch:** after local classification, the same merged-PR rows are matched\n   against each remote-tracking tip under the selected remote. When `headRefOid` equals that tip\n   and the branch is not the default, probe live existence with\n   `git ls-remote --heads \u003Cremote> refs\u002Fheads\u002F\u003Cbranch>`. A matching tip → `HIGH`\n   `merged-remote-branch` (remote head still present after merge, unset or blocked\n   `delete_branch_on_merge`). ls-remote failure → `MEDIUM` cached observation (may be stale after a\n   prune-less fetch). Empty ls-remote → no finding (head already gone upstream). Unlike every\n   other probe, ls-remote keeps global and system Git config, where the transport lives\n   (`core.sshCommand`, `credential.helper`, proxies). It runs only when the remote URL names the\n   same github.com repository with and without that config; otherwise it counts as a failure.\n   When every attempted probe in the run fails (non-zero status; empty success does not count),\n   emit one fleet-level `UNKNOWN` `ls-remote-fleet-unavailable` finding and withhold the\n   per-repository `MEDIUM` rows, because no live probe succeeded.\n   Remote-only heads (local already deleted) are included. The handoff is an optional `git push --delete --dry-run`\n   preview naming the remote and branch; this skill never runs it and never calls org-admin APIs to\n   flip repository settings. Enabling `delete_branch_on_merge` is complementary (it stops the class\n   accruing) and is **not** a substitute for this fleet visibility.\n   When the merged PR's `headRefOid` differs from the remote-tracking tip, the tip is still merged\n   content if it is an ancestor of `headRefOid` and that commit is in the clone\n   (`git merge-base --is-ancestor \u003Ctip> \u003CheadRefOid>` exits 0). That case is `MEDIUM`\n   `merged-remote-branch` with the reason `tip is an ancestor of the merged head`. It runs no\n   ls-remote probe, so it is never `HIGH`. A head commit absent from the clone, or a tip that is not\n   an ancestor, emits nothing.\n   **Remote branch families:** each remote-tracking branch under the selected remote whose name\n   belongs to a family below, and that no `merged-remote-branch` finding covers, gets one `LOW`\n   `remote-branch-family` row: the family, the tip, its age in days from the committer date, and\n   whether the tip is an ancestor of `\u003Cremote>\u002F\u003Cdefault>` (`yes`, `no`, or `unknown`). The row reads\n   the last-fetched `refs\u002Fremotes\u002F\u003Cremote>\u002F` inventory and adds no remote probe. It is report-only:\n   no handoff, no deletion, no deletion preview. Whether and when such branches are deleted is the\n   owner's decision.\n\n   | Family | Branch name | Source |\n   |---|---|---|\n   | `agent` | `agent-` followed by hex digits only | Claude Code subagent worktrees |\n   | `claude` | `claude\u002F*` | Claude Code cloud sessions (on the web, routine runs) |\n   | `plan` | `plan\u002F*` | named by hand |\n   | `stranded` | `stranded\u002F*` | named by hand |\n   | `pre-wipe` | `pre-wipe\u002F*` | ad hoc safety pushes made before a reimage |\n\n   | Claim | Basis | As of | Recheck |\n   |---|---|---|---|\n   | `agent-\u003Chex>` branches come from Claude Code subagent worktrees. | `plugins\u002Fsource-control\u002Fscripts\u002Fworktree-create.sh` uses the harness-supplied name verbatim as the branch. | 2026-09-29 | The harness stops supplying `agent-\u003Chex>` names, or that script changes how it names the branch. |\n   | `claude\u002F*` branches come from Claude Code cloud sessions. | Observed on fleet remotes (no probe recorded). Pointer: the routine-run branch prefix is in \u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Froutines#repositories-and-branch-permissions>; \u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fclaude-code-on-the-web> names none. | 2026-10-11 | That routines section changes the prefix or is renamed, the web page names a prefix, or a `claude\u002F*` branch turns up with another origin. |\n\n   **Unmerged remote branch:** a remote-tracking head with no `MERGED` PR row (not the default,\n   current, or worktree-attached branch) gets one more read-only GraphQL query for pull requests\n   whose `headRefName` is exactly that branch, in any state, plus `totalCount`. With a live\n   `ls-remote` tip: no PR at all and a complete list → `HIGH` `unmerged-remote-branch`, class\n   `never-pr`; only `CLOSED` PRs with one `headRefOid` equal to the live tip → `HIGH`, class\n   `closed-unmerged`. An `OPEN` or `MERGED` PR emits nothing. `ls-remote` failing → `MEDIUM`\n   `unmerged-remote-branch-unverified` (no tip, no deletion candidate). Only `CLOSED` PRs and none at\n   the live tip (commits pushed after close), or a `totalCount` above the returned page → `MEDIUM`\n   `unmerged-remote-branch-review`. Only the two `HIGH` classes reach the action plan, as\n   `delete-remote-branches` rows carrying the live tip as `expected_oid`, the remote's `remote_key`,\n   and the `github_repo` that apply rechecks. A branch whose only PR is merged stays\n   `merged-remote-branch` and is never a deletion candidate here.\n\n5. **Local inventories:** parse only `git worktree list --porcelain -z` registrations and\n   NUL-delimited `git for-each-ref` branch\u002Ftip records. Directory naming is\n   never worktree evidence. Compare each existing registered path's actual `--git-common-dir` with\n   the canonical checkout's expected common dir. A mismatch is `HIGH` evidence of an administrative\n   linkage problem but **manual review only**. Missing\u002Fprunable registrations never trigger pruning.\n   Linked, unlocked registrations with reliable admin emit one `MEDIUM` `worktree-status-handoff`\n   per repository naming those paths. Disposability (stranded \u002F unknown \u002F safe) is owned by\n   `\u002Fsource-control:worktree status`, and this collector emits no `git status`-based substitute\n   verdict. Separately, every linked worktree that passes existence and root-verifiability checks\n   is classified against its own repository's configured worktree root (`worktreeroot.path` read\n   with `git -C \u003Ccanonical>`, which honors `includeIf`, else the project directory's key, then\n   source-control `worktree_root`):\n   conforming, outside\u002Fwrong-layout\n   (expected `\u003Croot>\u002F\u003Cowner>-\u003Crepo>-\u003Cslug>` or `\u003Croot>\u002F\u003Crepo>-\u003Cslug>` without origin, matching\n   `\u002Fsource-control:worktree create`; create-shaped basenames stay conforming after branch\n   rename\u002Fdetach; comparisons use physical paths so symlink aliases of the configured root do not\n   false-positive), or tool-owned (Codex\u002FCursor).\n   Missing, prunable, non-root, and root-unverifiable registrations keep their own finding kinds and\n   are excluded from conformance denominators. When no root is configured, placement is reported\n   without asserting a convention. A repository with no `worktreeroot.path` of its own falls back\n   to the project directory's key, then pluginConfigs; it never borrows another target's root. The\n   header names the first resolvable root. The fleet summary lists each distinct root with its\n   repository count when repositories resolve different roots, and lists each repository whose own\n   root differed, with the root it used. Advice for a repository's worktrees names that\n   repository's root. If pluginConfigs cannot be read because `jq` is missing, emit\n   `worktree-root-pluginconfigs-unreadable` rather than pretending the key is unset.\n   Per-repository and fleet rollups always state the classifiable counts. If either inventory\n   command fails or emits malformed\u002Fpartial output, discard it, emit `UNKNOWN`, stop local\n   branch\u002Fworktree classification, and do not count that repository as successfully audited; an\n   empty\u002Ffailed inventory never means no branches are attached.\n6. **Protection:** current\u002Fdefault\u002Fworktree-attached branches are never emitted as standalone branch\n   cleanup candidates. A merged worktree is routed to worktree dry-run first. `merged-remote-branch`\n   is independent of local attachment. It describes the remote ref.\n\nEvery emitted finding kind, both confidence axes, and the merge-strategy and\n`gc.worktreePruneExpire` dependencies the tiers rest on:\n[reference\u002Fconfidence-model.md](reference\u002Fconfidence-model.md). The official Git\u002FGitHub behaviors\nthis collector relies on: [reference\u002Fofficial-sources.md](reference\u002Fofficial-sources.md). The\nread-only enforcement model and its threat assumptions:\n[reference\u002Fsecurity-review.md](reference\u002Fsecurity-review.md).\n\n## Presentation\n\nDefault output is screen-scale:\n\n1. Fleet header (config, scope, discovery counts).\n2. **Repository rollup**. One row per repository with `CLEAN` \u002F `N candidates` \u002F\n   `BLOCKED (evidence gap)`, plus counts by finding kind. Every UNKNOWN kind except the\n   disclosure-only `discovery-skip` and `discovery-symlink-skip` makes a repository, and so the fleet,\n   BLOCKED. Fleet-level findings (stale config,\n   duplicate checkouts) get their own row. A fleet verdict summarizes blocked vs candidate vs clean.\n3. **Fleet action plan**. Recommended skill invocations **once per repository** (not once per\n   finding), ordered so branch cleanups precede worktree cleanups, behind **one** confirmation gate.\n4. Path to the machine-readable action-plan JSON (and the `--apply-plan` dry-run invocation).\n\nPass `--detail` when the operator needs evidence: targets are collapsed (one entry per path\u002Fbranch\ncarrying every applicable finding), never duplicated across confidence groups. Never collapse\nsame-named branches across repositories.\n\n`ACKNOWLEDGED` is a prominence demotion, not a fifth confidence tier: the evidence stays exactly\nas weak as the `UNKNOWN` it came from. A rollup `CLEAN` verdict means no actionable cleanup-plan\ncandidates (the kinds that produce skill invocations) and no UNKNOWN evidence gap for that\nrepository (`discovery-skip` and `discovery-symlink-skip` are disclosure-only and do not count), not\n\"GitHub was unreachable so nothing was wrong.\" Manual-review HIGH\u002FMEDIUM findings\n(for example `locked-worktree` or `merged-pr-tip-drift`) remain in kind counts but do not inflate\n`N candidates` when the action plan correctly lists `Actions: none`.\n\nWhen the user wants to act on a fleet report, tell them to run\n`\u002Frepo-fleet-hygiene:apply --plan-file \u003Cpath>` (dry-run, then `--apply`) rather than driving\nper-repository skills by hand.\n\n## Fleet cleanup plan\n\nThis section defines how audit relates to the execute verb; it does not add a mutation command to\nthe audit argument grammar.\n\nThe cleanup-plan consumer is `\u002Frepo-fleet-hygiene:apply`. It takes only the machine-readable\nrollup artifact. Never parse this skill's human report into executable operations. The apply\nverb:\n\n1. rejects an incomplete, invalid, or non-audit artifact and preserves every repository-qualified\n   target, confidence, evidence gap, and disposition;\n2. owns batched merged-local-branch deletion (with fail-closed OID refresh) and worktree cleanup in\n   plan order rather than widening this audit script;\n3. presents one fleet action plan and obtains one explicit confirmation (or `--yes`) before any\n   mutation; and\n4. re-derives mutable facts, including relevant branch\u002Fworktree OIDs, at execution time. An old\n   artifact is evidence, not authorization.\n\nThe rollup, the `--apply-plan` dry-run, and `\u002Frepo-fleet-hygiene:apply` all ship. Do not\ninvent `--cleanup-plan`, `--execute`, or report-and-execute behavior on `audit-fleet.sh`. Return the\nreport, rollup, and plan path; tell the user to run `\u002Frepo-fleet-hygiene:apply` for execution. A\n`HIGH` evidence tier is never itself permission to delete a branch or worktree.\n\nRelated fleet contracts that remain separate:\n\n- merged remote branches carry a distinct safety gate: this skill reports them, and their deletion\n  is not part of `\u002Frepo-fleet-hygiene:apply`. Only never-PR and closed-unmerged remote branches can\n  be deleted, and only through `\u002Frepo-fleet-hygiene:apply --remote-branches` (per-branch prompt,\n  never bypassed by `--yes`, tip recorded first).\n\n## Graceful degradation\n\n- Git missing or too old: stop before scanning and give the prerequisite error.\n- Invalid config SYNTAX, invalid override, or an invalid CLI-supplied `--repo`\u002F`--root` path: report\n  the exact invalid input and stop; never silently fall back.\n- No scope given and no rung of the no-scope ladder resolves (exit 3): stop, and relay the script's\n  remedy block verbatim. The operator did not choose that path, so the rejection alone is not\n  actionable.\n- A config-sourced `fleet.repo`\u002F`fleet.root` path that is missing or not a Git working tree degrades\n  per-entry, not per-run: the entry becomes an `UNKNOWN` `stale-config-entry` finding and the rest of\n  the fleet is still audited (deleting repositories right after an audit must not abort every\n  subsequent run until the config is edited).\n- A path discovered under `--root` that is unreadable or not a Git working tree (despite a `.git`\n  marker) degrades the same way: an `UNKNOWN` `discovery-skip` finding, header skip counts, and the\n  rest of the fleet is still audited. The finding is disclosure-only: it does not make the fleet\n  BLOCKED. An explicitly named `--repo` that is not a working tree still\n  hard-fails.\n- A directory that itself carries a `.git` marker (directory or file) is treated as a nested\n  repository: discovery `add_target`s it and **returns without descending into its children**. A\n  repository buried inside another repository's working tree therefore never appears as its own\n  audit target unless named explicitly via `--repo` \u002F `fleet.repo`.\n- A symlinked or junctioned intermediate directory under `--root` is not followed, but is disclosed\n  as an `UNKNOWN` `discovery-symlink-skip` finding and counted on the discovery-skips header line.\n  Like `discovery-skip`, it is disclosure-only and does not make the fleet BLOCKED. What a Windows\n  directory junction is:\n  [Creating directory junctions](https:\u002F\u002Fgitforwindows.org\u002Fsymbolic-links.html#creating-directory-junctions).\n  Symlinked\n  discovery *roots* remain a hard refusal (CLI) or `stale-config-entry` (configured).\n- `gh` missing\u002Funauthenticated or API\u002Ftimeout failure: continue Git\u002Fworktree checks, report GitHub\n  evidence as `UNKNOWN`, and make no merged\u002Fmigration claim. Compatible `timeout`\u002F`gtimeout` is\n  preferred; otherwise use the collector's finite TERM-to-KILL Bash watchdog.\n- Non-GitHub or ambiguous remote: continue local checks; GitHub identity\u002FPR evidence is `UNKNOWN`.\n- Canonical override is missing a GitHub remote or does not resolve to the same normalized GitHub\n  identity as the discovered repo: surface `UNKNOWN`, stop that repository, and do not merge evidence.\n- Worktree porcelain fails: surface `UNKNOWN` and stop local branch\u002Fworktree classification for that\n  repository so no branch can be mislabeled unattached.\n- Branch enumeration fails or is malformed: discard every partial record, surface `UNKNOWN`, stop\n  branch classification, and exclude the repository from the successful-audit count.\n\n## Integration\n\n| Finding | Handoff (not executed here) |\n|---|---|\n| `merged-local-branch` | Run `\u002Frepo-hygiene:clean git` in the named canonical repository |\n| `merged-remote-branch` | Optional preview only: `git push --delete --dry-run \u003Cremote> \u003Cbranch>` in the canonical repository (never executed here). Enabling GitHub `delete_branch_on_merge` is complementary and owned by the repository's settings automation. This audit does not change it |\n| `unmerged-remote-branch` | Gated delete: `apply-plan.sh --plan-file \u003Cpath> --apply --remote-branches` in the operator's own terminal, one prompt per branch, deleted only at the recorded tip; `\u002Frepo-fleet-hygiene:apply --remote-branches` from a session without a terminal only previews. Preview by hand: `git push --delete --dry-run \u003Cremote> \u003Cbranch>` in the canonical repository |\n| `unmerged-remote-branch-unverified` | Confirm `git ls-remote --heads` works by hand, then rerun. No deletion is planned |\n| `unmerged-remote-branch-review` | Manual review of the branch and its closed PRs; never a deletion candidate |\n| `remote-branch-family` | None. Report only: no deletion and no deletion preview. Read the family, age and on-default fields and decide by hand |\n| `merged-worktree`, `prunable-worktree`, `missing-worktree` | Run `\u002Fsource-control:worktree cleanup --dry-run` in the canonical repository |\n| `worktree-status-handoff` | Run `\u002Fsource-control:worktree status` in the canonical repository (stranded-work axis); use cleanup `--dry-run` only after Work is safe. If `source-control` is not installed, name the listed worktree targets and the missing collaborator. Emit no porcelain-based substitute verdict |\n| `worktree-admin-mismatch` | Manual inspection; `git worktree repair` is an option only after validating which administrative directory is authoritative |\n| `worktree-not-a-root` | Manual inspection of the registered path; a `git -C` probe of it describes the CONTAINING repository, so no cleanup handoff is safe until the path is resolved |\n| `worktree-root-unverifiable` | Manual inspection of the registered path. Root-ness is unproven here rather than disproven. The probe itself failed, so infer nothing about the path in either direction |\n| `worktree-nested-in-repository` | Recreate at an external root with `\u002Fsource-control:worktree create`, then remove the nested one |\n| `worktree-outside-configured-root` | Recreate at the expected location named in evidence with `\u002Fsource-control:worktree create`, then remove the misplaced one |\n| `worktree-wrong-layout` | Recreate at the expected `\u003Cowner>-\u003Crepo>-\u003Cslug>` path under the configured root, then remove the wrong-layout one |\n| `worktree-tool-owned` | Leave to Codex\u002FCursor lifecycle, or migrate deliberately to the configured root |\n| `worktree-root-conformance` | Read the per-worktree outside\u002Fwrong-layout findings for expected paths; migrate toward the configured root |\n| `worktree-root-conformance-summary` | Same as per-repository conformance; fleet-scale migration toward the configured root |\n| `worktree-root-unconfigured` | Set `worktreeroot.path` (git config) or source-control `worktree_root`, then rerun |\n| `ls-remote-fleet-unavailable` | Confirm `git ls-remote --heads` works by hand with the operator's usual Git transport, then rerun |\n| `worktree-root-pluginconfigs-unreadable` | Install `jq`, or set `worktreeroot.path`; do not treat the fleet as unconfigured |\n| `worktree-placement-unverifiable` | Inspect the canonical checkout; placement was not checked for any of its worktrees, so their placement is unknown rather than confirmed |\n| `bare-repo-with-working-tree` | Manual review: `core.bare=true` coincides with working-tree content or registered linked worktrees. What the setting does: the [`core.bare`](https:\u002F\u002Fgit-scm.com\u002Fdocs\u002Fgit-config#Documentation\u002Fgit-config.txt-corebare) entry |\n| `github-remote-moved` | Human-reviewed `git remote set-url`; this plugin never changes remotes |\n\nThis plugin remains useful if those optional collaborators are absent: the report names the local\nGit\u002FGitHub evidence and target so another tool or human can act. For `worktree-status-handoff`, when\n`source-control` is missing, name the listed worktree paths and the missing collaborator and do not\ninvent a porcelain-based disposability substitute.\n\n## Gotchas\n\n- **Which config is consumed depends on where the audit runs.** Config resolution follows the ladder in\n  the Input resolution section: explicit `--config`, else the project-scoped\n  `.claude\u002Frepo-fleet-hygiene.conf`, else the user-global one. A project-scoped config is invisible when\n  the audit runs from a different project, so confirm the consumed config named in the report header\n  before trusting a run's scope.\n- **Config paths resolve relative to the config file's directory.** A relative `root`\u002F`repo`\u002Fcanonical\n  path is anchored at the config directory, not the audit's working directory. Absolute paths work but\n  are what a consumer's write-time path guard flags, so author config via\n  `\u002Frepo-fleet-hygiene:setup apply`, which prefers the portable relative form.\n","**Arguments.** `[[\u003Cdir>]... [--root \u003Cdir>]... [--repo \u003Cdir>]... [--config \u003Cfile>] [--detail]|--apply-plan \u003Cpath>]`. Full form: [[\u003Cdir>]... [--root \u003Cdir>]... [--repo \u003Cdir>]... [--config \u003Cfile>] [--canonical \u003Cgithub.com\u002Fowner\u002Frepo=path>]... [--skip \u003Cname>]... [--extend-skip \u003Cname>]... [--max-depth \u003C1..12>] [--detail] [--plan-file \u003Cpath>]|--apply-plan \u003Cpath>]\n\n## Purpose\n\nCoordinate cross-repository hygiene. This skill owns bounded fleet discovery, canonical\ncheckout resolution, fleet-scale evidence collection, rollup, and action-plan routing. It does\n**not** own per-repository cleanup decisions or execution; those belong to `repo-hygiene` and\n`source-control`.\n\nThe collector produces the detailed read-only report described below, including the compact\nmachine-readable rollup and action-plan artifact. Tell the user to run\n`\u002Frepo-fleet-hygiene:apply --plan-file \u003Cpath>` (dry-run by default; `--apply` plus confirmation or\n`--yes` to mutate; `--remote-branches` additionally opts in to per-branch remote deletion, which\n`--yes` never answers). Do not add an execute flag to this audit script.\n\n## Non-negotiable boundary\n\nNever run or suggest running inline from this skill: `git fetch`, `git worktree prune`,\n`git worktree repair`, `git worktree remove`, `git branch -d\u002F-D`, `git remote set-url`, or any\nfilesystem deletion. The bundled script has no mutation mode. `--apply-plan` is a read-only\ndry-run approval artifact over a prior plan file. A report may name a command\u002Ftool as a future\nhandoff; the fleet action plan lists those invocations once per repository behind one confirmation\ngate. Actual fleet mutation belongs to `\u002Frepo-fleet-hygiene:apply`, not this skill.\n\n## Input resolution\n\nParse `$ARGUMENTS` as opaque arguments for the bundled script. Supported flags:\n\n- `\u003Cdir>`: a bare positional path, treated as `--root`. A drive root (`D:`, `D:\u002F`) is a legitimate\n  discovery root and normalizes to `D:\u002F`. This is the form `\u002Frepo-fleet-hygiene:audit D:` uses.\n- `--root \u003Cdir>`: bounded recursive repository discovery (repeatable).\n- `--repo \u003Cdir>`: exact repository\u002Fworktree target (repeatable).\n- `--config \u003Cfile>`: explicit Git-format config (at most one).\n- `--canonical \u003Cgithub.com\u002Fowner\u002Frepo=path>`: invocation-specific canonical checkout override\n  (repeatable; explicit wins over config).\n- `--skip \u003Cname>`: discovery directory-name skip (repeatable). Explicit `--skip` \u002F `fleet.skip`\n  entries **replace** the default skip list rather than appending. Otherwise shrinking is\n  impossible. Default (neither CLI nor config): `vendor` plus the package-manager cache trees\n  `node_modules`, `.venv`, `.pnpm-store`, `.yarn`, `.npm`, `.cargo`, `.rustup`, `.gradle`, `.m2`,\n  `.nuget`, `__pycache__`, and `.tox`. To shrink (e.g. reach a repo under `vendor\u002F`), list only the\n  names you still want skipped. CLI and config compose additively with each other like other scope\n  inputs. Values must be bare directory names (no empty value, no path separator). `.`, `..`, and\n  `.git` stay skipped unconditionally even when an explicit list omits them.\n- `--extend-skip \u003Cname>`: discovery directory-name skip that **adds** to whichever list is in\n  effect (the defaults, or an explicit `--skip` \u002F `fleet.skip` list) instead of replacing it\n  (repeatable). Config equivalent: repeatable `fleet.skipAppend`. Same bare-name validation.\n- `--max-depth \u003C1..12>`: discovery bound; explicit wins over config\u002Fdefault `5`.\n- `--project-dir \u003Cdir>`: the session's project directory, used for the project-scoped config rung.\n  It is **not** a scope fallback; see the no-scope ladder below.\n- `--detail`: emit collapsed per-target evidence after the rollup (default is rollup + action plan\n  only).\n- `--plan-file \u003Cpath>`: write the machine-readable action-plan JSON to this path (otherwise a temp\n  file is created and named in the report).\n- `--apply-plan \u003Cpath>`: standalone read-only mode. Render the ordered dry-run approval artifact\n  for a previously written plan (cannot combine with discovery flags).\n\nAlways pass `--project-dir \"${CLAUDE_PROJECT_DIR}\"` on audit runs (not on `--apply-plan`). The\nscript does not read the project directory from its environment, so passing it in is what makes\nthe project rung below reachable at all.\n\n- **Pointer**: where the token is substituted is in\n  [Available string substitutions](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fskills#available-string-substitutions),\n  and which processes receive the variable is in\n  [Where each variable resolves](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmanifest-reference#where-each-variable-resolves).\n\nIf no explicit scope and no config-supplied `fleet.root`\u002F`fleet.repo` resolve, the run uses the\nshared no-scope ladder: `--named` paths, then `ghq root --all` when `ghq` is installed, then the\nnearest of the 4 parents above the working directory's checkout (above the working directory itself\noutside a checkout) that directly holds 2 or more Git repositories, so a run inside one checkout\ncovers its sibling fleet, then the checkout holding the working directory, else exit 3 naming every\nrung. A Git checkout in the working directory is a rung; the session project directory is not one\non its own. **Claim:** `ghq root --all` prints every configured root. **Basis:** the\n[ghq README](https:\u002F\u002Fgithub.com\u002Fx-motemen\u002Fghq#commands) (`ghq root [--all]`; \"Without '--all' option,\nthe primary one is shown\") and `ghq help root` on ghq 1.10.1 (\"--all  Show all roots\"). **As of:**\n2026-11-10. **Recheck:** the README's `root` entry drops or renames `--all`. Pass that guidance\nthrough rather than re-deriving a root yourself. Config resolution is the script's own ladder. Do not\npre-resolve or pass a probed path yourself:\nexplicit `--config` wins, else the script probes\n`\u003Cproject-dir>\u002F.claude\u002Frepo-fleet-hygiene.conf` (project-scoped), else\n`~\u002F.claude\u002Frepo-fleet-hygiene.conf` (user-global, a machine-scoped fleet config placed there is\nrecorded user intent, not a guessed root). The report header names the consumed config and its\nsource, or states that none was consumed. Never guess a broader machine root from the current\npath beyond that ladder.\n\nConfig-supplied scope is **additive** to CLI-supplied scope: a `--repo X` run still walks every\nconfigured root. The header's `Scope:` line names each contributing rung and its entry count, so\nreport that line rather than assuming the arguments were the whole scope.\n\nBefore execution, reject any arguments outside this grammar, noting that a bare positional path\n**is** in the grammar, so `\u002Frepo-fleet-hygiene:audit D:` and\n`\u002Frepo-fleet-hygiene:audit \u002Fpath\u002Fto\u002Ftree` are valid invocations to pass through, not arguments to\nrefuse. What stays rejected is an unrecognized flag: anything beginning with `-` that is not listed\nabove. Pass every path\u002Foverride as a quoted argument; never assemble a shell fragment from config,\nrepository, remote, or branch text.\n\nRun exactly once:\n\n```bash\n${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-fleet.sh --project-dir \"${CLAUDE_PROJECT_DIR}\" \u003Cvalidated-and-quoted-arguments>\n```\n\nThe script validates config with `git config --file`; it never sources or executes it.\n\n## Evidence rules\n\nThe bundled collector is authoritative for classifications. Preserve its evidence in the report:\n\n1. **Canonical checkout:** explicit remote-keyed override → configured remote-keyed override → the\n   repository's **main worktree**, read from `git worktree list --porcelain`, never from the\n   checkout discovery reached first. Which record is the main worktree: git-worktree's\n   [`list`](https:\u002F\u002Fgit-scm.com\u002Fdocs\u002Fgit-worktree#Documentation\u002Fgit-worktree.txt-list) entry and\n   [List output format](https:\u002F\u002Fgit-scm.com\u002Fdocs\u002Fgit-worktree#_list_output_format). When a supplied or discovered path resolves to a different main\n   worktree, the header states the substitution on one `Resolved to main worktree:` line per\n   repository, naming every path that resolved into it. Relay it, because the operator named one\n   path and the report is about another. The report always shows\n   discovered and canonical paths. An\n   override target with a missing\u002Fnon-GitHub remote, or a different identity that cannot be proven to\n   resolve to the same GitHub repository, stops that repository's local audit before evidence combines.\n2. **GitHub identity:** read the selected fetch remote with `git remote get-url`; accept only\n   `github.com\u002Fowner\u002Frepo`; query `GET \u002Frepos\u002F{owner}\u002F{repo}`. If returned `full_name` differs, report\n   `HIGH` transfer\u002Frename evidence and **continue** branch\u002Fworktree analysis against that resolved\n   identity. A moved remote is not a reason to skip local classification or merge evidence. A\n   404\u002F403\u002Fnetwork error is `UNKNOWN`, never \"deleted\" or \"moved\". A 404\u002F403 on an identity listed\n   in `fleet.ackUnavailable` is demoted to `ACKNOWLEDGED`, still reported, never suppressed; acks\n   never touch non-404\u002F403 failures or successful-response evidence.\n3. **Merged branch:** one aliased `gh api graphql` query per repository page of local branches\n   (up to `MERGED_PR_GRAPHQL_ALIAS_PAGE` `headRefName` aliases per call, `first:1`,\n   `states:[MERGED]`). GraphQL's `headRefName` argument is an **exact** match, never the search\n   API's prefix-matching `head:` qualifier, so `feature\u002Fauth` and `feature\u002Fauth-v2` never conflate.\n   Measured rate cost stays 1 per call (nodeCount equals the alias count), well under GitHub's\n   limits:\n   [Primary rate limit](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fgraphql\u002Foverview\u002Frate-limits-and-query-limits-for-the-graphql-api#primary-rate-limit)\n   and [Node limit](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fgraphql\u002Foverview\u002Frate-limits-and-query-limits-for-the-graphql-api#node-limit),\n   as of 2026-11-03; recheck when either section changes its limit. There is no merged-PR\n   result window and no per-branch fallback: every non-default local branch the operator asked\n   about is queried by exact name, including heads GitHub auto-deleted and a later fetch pruned. Fail closed when `gh`\u002FGraphQL is\n   unavailable. Emit `github-pr-evidence-unavailable` and never infer unmerged from a missing\n   row after a failed page. Identical branch names in another repository are unrelated. `HIGH`\n   requires the PR `headRefOid` to equal the current local tip. Tip drift is `MEDIUM` manual\n   review. Git ancestry without GitHub evidence is `LOW` and never called merged-by-PR, and\n   under squash merges that ancestry predicate is near-inert, so on a squash-merging fleet\n   GitHub evidence is effectively the only merge evidence.\n4. **Merged remote branch:** after local classification, the same merged-PR rows are matched\n   against each remote-tracking tip under the selected remote. When `headRefOid` equals that tip\n   and the branch is not the default, probe live existence with\n   `git ls-remote --heads \u003Cremote> refs\u002Fheads\u002F\u003Cbranch>`. A matching tip → `HIGH`\n   `merged-remote-branch` (remote head still present after merge, unset or blocked\n   `delete_branch_on_merge`). ls-remote failure → `MEDIUM` cached observation (may be stale after a\n   prune-less fetch). Empty ls-remote → no finding (head already gone upstream). Unlike every\n   other probe, ls-remote keeps global and system Git config, where the transport lives\n   (`core.sshCommand`, `credential.helper`, proxies). It runs only when the remote URL names the\n   same github.com repository with and without that config; otherwise it counts as a failure.\n   When every attempted probe in the run fails (non-zero status; empty success does not count),\n   emit one fleet-level `UNKNOWN` `ls-remote-fleet-unavailable` finding and withhold the\n   per-repository `MEDIUM` rows, because no live probe succeeded.\n   Remote-only heads (local already deleted) are included. The handoff is an optional `git push --delete --dry-run`\n   preview naming the remote and branch; this skill never runs it and never calls org-admin APIs to\n   flip repository settings. Enabling `delete_branch_on_merge` is complementary (it stops the class\n   accruing) and is **not** a substitute for this fleet visibility.\n   When the merged PR's `headRefOid` differs from the remote-tracking tip, the tip is still merged\n   content if it is an ancestor of `headRefOid` and that commit is in the clone\n   (`git merge-base --is-ancestor \u003Ctip> \u003CheadRefOid>` exits 0). That case is `MEDIUM`\n   `merged-remote-branch` with the reason `tip is an ancestor of the merged head`. It runs no\n   ls-remote probe, so it is never `HIGH`. A head commit absent from the clone, or a tip that is not\n   an ancestor, emits nothing.\n   **Remote branch families:** each remote-tracking branch under the selected remote whose name\n   belongs to a family below, and that no `merged-remote-branch` finding covers, gets one `LOW`\n   `remote-branch-family` row: the family, the tip, its age in days from the committer date, and\n   whether the tip is an ancestor of `\u003Cremote>\u002F\u003Cdefault>` (`yes`, `no`, or `unknown`). The row reads\n   the last-fetched `refs\u002Fremotes\u002F\u003Cremote>\u002F` inventory and adds no remote probe. It is report-only:\n   no handoff, no deletion, no deletion preview. Whether and when such branches are deleted is the\n   owner's decision.\n\n   | Family | Branch name | Source |\n   |---|---|---|\n   | `agent` | `agent-` followed by hex digits only | Claude Code subagent worktrees |\n   | `claude` | `claude\u002F*` | Claude Code cloud sessions (on the web, routine runs) |\n   | `plan` | `plan\u002F*` | named by hand |\n   | `stranded` | `stranded\u002F*` | named by hand |\n   | `pre-wipe` | `pre-wipe\u002F*` | ad hoc safety pushes made before a reimage |\n\n   | Claim | Basis | As of | Recheck |\n   |---|---|---|---|\n   | `agent-\u003Chex>` branches come from Claude Code subagent worktrees. | `plugins\u002Fsource-control\u002Fscripts\u002Fworktree-create.sh` uses the harness-supplied name verbatim as the branch. | 2026-09-29 | The harness stops supplying `agent-\u003Chex>` names, or that script changes how it names the branch. |\n   | `claude\u002F*` branches come from Claude Code cloud sessions. | Observed on fleet remotes (no probe recorded). Pointer: the routine-run branch prefix is in \u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Froutines#repositories-and-branch-permissions>; \u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fclaude-code-on-the-web> names none. | 2026-10-11 | That routines section changes the prefix or is renamed, the web page names a prefix, or a `claude\u002F*` branch turns up with another origin. |\n\n   **Unmerged remote branch:** a remote-tracking head with no `MERGED` PR row (not the default,\n   current, or worktree-attached branch) gets one more read-only GraphQL query for pull requests\n   whose `headRefName` is exactly that branch, in any state, plus `totalCount`. With a live\n   `ls-remote` tip: no PR at all and a complete list → `HIGH` `unmerged-remote-branch`, class\n   `never-pr`; only `CLOSED` PRs with one `headRefOid` equal to the live tip → `HIGH`, class\n   `closed-unmerged`. An `OPEN` or `MERGED` PR emits nothing. `ls-remote` failing → `MEDIUM`\n   `unmerged-remote-branch-unverified` (no tip, no deletion candidate). Only `CLOSED` PRs and none at\n   the live tip (commits pushed after close), or a `totalCount` above the returned page → `MEDIUM`\n   `unmerged-remote-branch-review`. Only the two `HIGH` classes reach the action plan, as\n   `delete-remote-branches` rows carrying the live tip as `expected_oid`, the remote's `remote_key`,\n   and the `github_repo` that apply rechecks. A branch whose only PR is merged stays\n   `merged-remote-branch` and is never a deletion candidate here.\n\n5. **Local inventories:** parse only `git worktree list --porcelain -z` registrations and\n   NUL-delimited `git for-each-ref` branch\u002Ftip records. Directory naming is\n   never worktree evidence. Compare each existing registered path's actual `--git-common-dir` with\n   the canonical checkout's expected common dir. A mismatch is `HIGH` evidence of an administrative\n   linkage problem but **manual review only**. Missing\u002Fprunable registrations never trigger pruning.\n   Linked, unlocked registrations with reliable admin emit one `MEDIUM` `worktree-status-handoff`\n   per repository naming those paths. Disposability (stranded \u002F unknown \u002F safe) is owned by\n   `\u002Fsource-control:worktree status`, and this collector emits no `git status`-based substitute\n   verdict. Separately, every linked worktree that passes existence and root-verifiability checks\n   is classified against its own repository's configured worktree root (`worktreeroot.path` read\n   with `git -C \u003Ccanonical>`, which honors `includeIf`, else the project directory's key, then\n   source-control `worktree_root`):\n   conforming, outside\u002Fwrong-layout\n   (expected `\u003Croot>\u002F\u003Cowner>-\u003Crepo>-\u003Cslug>` or `\u003Croot>\u002F\u003Crepo>-\u003Cslug>` without origin, matching\n   `\u002Fsource-control:worktree create`; create-shaped basenames stay conforming after branch\n   rename\u002Fdetach; comparisons use physical paths so symlink aliases of the configured root do not\n   false-positive), or tool-owned (Codex\u002FCursor).\n   Missing, prunable, non-root, and root-unverifiable registrations keep their own finding kinds and\n   are excluded from conformance denominators. When no root is configured, placement is reported\n   without asserting a convention. A repository with no `worktreeroot.path` of its own falls back\n   to the project directory's key, then pluginConfigs; it never borrows another target's root. The\n   header names the first resolvable root. The fleet summary lists each distinct root with its\n   repository count when repositories resolve different roots, and lists each repository whose own\n   root differed, with the root it used. Advice for a repository's worktrees names that\n   repository's root. If pluginConfigs cannot be read because `jq` is missing, emit\n   `worktree-root-pluginconfigs-unreadable` rather than pretending the key is unset.\n   Per-repository and fleet rollups always state the classifiable counts. If either inventory\n   command fails or emits malformed\u002Fpartial output, discard it, emit `UNKNOWN`, stop local\n   branch\u002Fworktree classification, and do not count that repository as successfully audited; an\n   empty\u002Ffailed inventory never means no branches are attached.\n6. **Protection:** current\u002Fdefault\u002Fworktree-attached branches are never emitted as standalone branch\n   cleanup candidates. A merged worktree is routed to worktree dry-run first. `merged-remote-branch`\n   is independent of local attachment. It describes the remote ref.\n\nEvery emitted finding kind, both confidence axes, and the merge-strategy and\n`gc.worktreePruneExpire` dependencies the tiers rest on:\n[reference\u002Fconfidence-model.md](reference\u002Fconfidence-model.md). The official Git\u002FGitHub behaviors\nthis collector relies on: [reference\u002Fofficial-sources.md](reference\u002Fofficial-sources.md). The\nread-only enforcement model and its threat assumptions:\n[reference\u002Fsecurity-review.md](reference\u002Fsecurity-review.md).\n\n## Presentation\n\nDefault output is screen-scale:\n\n1. Fleet header (config, scope, discovery counts).\n2. **Repository rollup**. One row per repository with `CLEAN` \u002F `N candidates` \u002F\n   `BLOCKED (evidence gap)`, plus counts by finding kind. Every UNKNOWN kind except the\n   disclosure-only `discovery-skip` and `discovery-symlink-skip` makes a repository, and so the fleet,\n   BLOCKED. Fleet-level findings (stale config,\n   duplicate checkouts) get their own row. A fleet verdict summarizes blocked vs candidate vs clean.\n3. **Fleet action plan**. Recommended skill invocations **once per repository** (not once per\n   finding), ordered so branch cleanups precede worktree cleanups, behind **one** confirmation gate.\n4. Path to the machine-readable action-plan JSON (and the `--apply-plan` dry-run invocation).\n\nPass `--detail` when the operator needs evidence: targets are collapsed (one entry per path\u002Fbranch\ncarrying every applicable finding), never duplicated across confidence groups. Never collapse\nsame-named branches across repositories.\n\n`ACKNOWLEDGED` is a prominence demotion, not a fifth confidence tier: the evidence stays exactly\nas weak as the `UNKNOWN` it came from. A rollup `CLEAN` verdict means no actionable cleanup-plan\ncandidates (the kinds that produce skill invocations) and no UNKNOWN evidence gap for that\nrepository (`discovery-skip` and `discovery-symlink-skip` are disclosure-only and do not count), not\n\"GitHub was unreachable so nothing was wrong.\" Manual-review HIGH\u002FMEDIUM findings\n(for example `locked-worktree` or `merged-pr-tip-drift`) remain in kind counts but do not inflate\n`N candidates` when the action plan correctly lists `Actions: none`.\n\nWhen the user wants to act on a fleet report, tell them to run\n`\u002Frepo-fleet-hygiene:apply --plan-file \u003Cpath>` (dry-run, then `--apply`) rather than driving\nper-repository skills by hand.\n\n## Fleet cleanup plan\n\nThis section defines how audit relates to the execute verb; it does not add a mutation command to\nthe audit argument grammar.\n\nThe cleanup-plan consumer is `\u002Frepo-fleet-hygiene:apply`. It takes only the machine-readable\nrollup artifact. Never parse this skill's human report into executable operations. The apply\nverb:\n\n1. rejects an incomplete, invalid, or non-audit artifact and preserves every repository-qualified\n   target, confidence, evidence gap, and disposition;\n2. owns batched merged-local-branch deletion (with fail-closed OID refresh) and worktree cleanup in\n   plan order rather than widening this audit script;\n3. presents one fleet action plan and obtains one explicit confirmation (or `--yes`) before any\n   mutation; and\n4. re-derives mutable facts, including relevant branch\u002Fworktree OIDs, at execution time. An old\n   artifact is evidence, not authorization.\n\nThe rollup, the `--apply-plan` dry-run, and `\u002Frepo-fleet-hygiene:apply` all ship. Do not\ninvent `--cleanup-plan`, `--execute`, or report-and-execute behavior on `audit-fleet.sh`. Return the\nreport, rollup, and plan path; tell the user to run `\u002Frepo-fleet-hygiene:apply` for execution. A\n`HIGH` evidence tier is never itself permission to delete a branch or worktree.\n\nRelated fleet contracts that remain separate:\n\n- merged remote branches carry a distinct safety gate: this skill reports them, and their deletion\n  is not part of `\u002Frepo-fleet-hygiene:apply`. Only never-PR and closed-unmerged remote branches can\n  be deleted, and only through `\u002Frepo-fleet-hygiene:apply --remote-branches` (per-branch prompt,\n  never bypassed by `--yes`, tip recorded first).\n\n## Graceful degradation\n\n- Git missing or too old: stop before scanning and give the prerequisite error.\n- Invalid config SYNTAX, invalid override, or an invalid CLI-supplied `--repo`\u002F`--root` path: report\n  the exact invalid input and stop; never silently fall back.\n- No scope given and no rung of the no-scope ladder resolves (exit 3): stop, and relay the script's\n  remedy block verbatim. The operator did not choose that path, so the rejection alone is not\n  actionable.\n- A config-sourced `fleet.repo`\u002F`fleet.root` path that is missing or not a Git working tree degrades\n  per-entry, not per-run: the entry becomes an `UNKNOWN` `stale-config-entry` finding and the rest of\n  the fleet is still audited (deleting repositories right after an audit must not abort every\n  subsequent run until the config is edited).\n- A path discovered under `--root` that is unreadable or not a Git working tree (despite a `.git`\n  marker) degrades the same way: an `UNKNOWN` `discovery-skip` finding, header skip counts, and the\n  rest of the fleet is still audited. The finding is disclosure-only: it does not make the fleet\n  BLOCKED. An explicitly named `--repo` that is not a working tree still\n  hard-fails.\n- A directory that itself carries a `.git` marker (directory or file) is treated as a nested\n  repository: discovery `add_target`s it and **returns without descending into its children**. A\n  repository buried inside another repository's working tree therefore never appears as its own\n  audit target unless named explicitly via `--repo` \u002F `fleet.repo`.\n- A symlinked or junctioned intermediate directory under `--root` is not followed, but is disclosed\n  as an `UNKNOWN` `discovery-symlink-skip` finding and counted on the discovery-skips header line.\n  Like `discovery-skip`, it is disclosure-only and does not make the fleet BLOCKED. What a Windows\n  directory junction is:\n  [Creating directory junctions](https:\u002F\u002Fgitforwindows.org\u002Fsymbolic-links.html#creating-directory-junctions).\n  Symlinked\n  discovery *roots* remain a hard refusal (CLI) or `stale-config-entry` (configured).\n- `gh` missing\u002Funauthenticated or API\u002Ftimeout failure: continue Git\u002Fworktree checks, report GitHub\n  evidence as `UNKNOWN`, and make no merged\u002Fmigration claim. Compatible `timeout`\u002F`gtimeout` is\n  preferred; otherwise use the collector's finite TERM-to-KILL Bash watchdog.\n- Non-GitHub or ambiguous remote: continue local checks; GitHub identity\u002FPR evidence is `UNKNOWN`.\n- Canonical override is missing a GitHub remote or does not resolve to the same normalized GitHub\n  identity as the discovered repo: surface `UNKNOWN`, stop that repository, and do not merge evidence.\n- Worktree porcelain fails: surface `UNKNOWN` and stop local branch\u002Fworktree classification for that\n  repository so no branch can be mislabeled unattached.\n- Branch enumeration fails or is malformed: discard every partial record, surface `UNKNOWN`, stop\n  branch classification, and exclude the repository from the successful-audit count.\n\n## Integration\n\n| Finding | Handoff (not executed here) |\n|---|---|\n| `merged-local-branch` | Run `\u002Frepo-hygiene:clean git` in the named canonical repository |\n| `merged-remote-branch` | Optional preview only: `git push --delete --dry-run \u003Cremote> \u003Cbranch>` in the canonical repository (never executed here). Enabling GitHub `delete_branch_on_merge` is complementary and owned by the repository's settings automation. This audit does not change it |\n| `unmerged-remote-branch` | Gated delete: `apply-plan.sh --plan-file \u003Cpath> --apply --remote-branches` in the operator's own terminal, one prompt per branch, deleted only at the recorded tip; `\u002Frepo-fleet-hygiene:apply --remote-branches` from a session without a terminal only previews. Preview by hand: `git push --delete --dry-run \u003Cremote> \u003Cbranch>` in the canonical repository |\n| `unmerged-remote-branch-unverified` | Confirm `git ls-remote --heads` works by hand, then rerun. No deletion is planned |\n| `unmerged-remote-branch-review` | Manual review of the branch and its closed PRs; never a deletion candidate |\n| `remote-branch-family` | None. Report only: no deletion and no deletion preview. Read the family, age and on-default fields and decide by hand |\n| `merged-worktree`, `prunable-worktree`, `missing-worktree` | Run `\u002Fsource-control:worktree cleanup --dry-run` in the canonical repository |\n| `worktree-status-handoff` | Run `\u002Fsource-control:worktree status` in the canonical repository (stranded-work axis); use cleanup `--dry-run` only after Work is safe. If `source-control` is not installed, name the listed worktree targets and the missing collaborator. Emit no porcelain-based substitute verdict |\n| `worktree-admin-mismatch` | Manual inspection; `git worktree repair` is an option only after validating which administrative directory is authoritative |\n| `worktree-not-a-root` | Manual inspection of the registered path; a `git -C` probe of it describes the CONTAINING repository, so no cleanup handoff is safe until the path is resolved |\n| `worktree-root-unverifiable` | Manual inspection of the registered path. Root-ness is unproven here rather than disproven. The probe itself failed, so infer nothing about the path in either direction |\n| `worktree-nested-in-repository` | Recreate at an external root with `\u002Fsource-control:worktree create`, then remove the nested one |\n| `worktree-outside-configured-root` | Recreate at the expected location named in evidence with `\u002Fsource-control:worktree create`, then remove the misplaced one |\n| `worktree-wrong-layout` | Recreate at the expected `\u003Cowner>-\u003Crepo>-\u003Cslug>` path under the configured root, then remove the wrong-layout one |\n| `worktree-tool-owned` | Leave to Codex\u002FCursor lifecycle, or migrate deliberately to the configured root |\n| `worktree-root-conformance` | Read the per-worktree outside\u002Fwrong-layout findings for expected paths; migrate toward the configured root |\n| `worktree-root-conformance-summary` | Same as per-repository conformance; fleet-scale migration toward the configured root |\n| `worktree-root-unconfigured` | Set `worktreeroot.path` (git config) or source-control `worktree_root`, then rerun |\n| `ls-remote-fleet-unavailable` | Confirm `git ls-remote --heads` works by hand with the operator's usual Git transport, then rerun |\n| `worktree-root-pluginconfigs-unreadable` | Install `jq`, or set `worktreeroot.path`; do not treat the fleet as unconfigured |\n| `worktree-placement-unverifiable` | Inspect the canonical checkout; placement was not checked for any of its worktrees, so their placement is unknown rather than confirmed |\n| `bare-repo-with-working-tree` | Manual review: `core.bare=true` coincides with working-tree content or registered linked worktrees. What the setting does: the [`core.bare`](https:\u002F\u002Fgit-scm.com\u002Fdocs\u002Fgit-config#Documentation\u002Fgit-config.txt-corebare) entry |\n| `github-remote-moved` | Human-reviewed `git remote set-url`; this plugin never changes remotes |\n\nThis plugin remains useful if those optional collaborators are absent: the report names the local\nGit\u002FGitHub evidence and target so another tool or human can act. For `worktree-status-handoff`, when\n`source-control` is missing, name the listed worktree paths and the missing collaborator and do not\ninvent a porcelain-based disposability substitute.\n\n## Gotchas\n\n- **Which config is consumed depends on where the audit runs.** Config resolution follows the ladder in\n  the Input resolution section: explicit `--config`, else the project-scoped\n  `.claude\u002Frepo-fleet-hygiene.conf`, else the user-global one. A project-scoped config is invisible when\n  the audit runs from a different project, so confirm the consumed config named in the report header\n  before trusting a run's scope.\n- **Config paths resolve relative to the config file's directory.** A relative `root`\u002F`repo`\u002Fcanonical\n  path is anchored at the config directory, not the audit's working directory. Absolute paths work but\n  are what a consumer's write-time path guard flags, so author config via\n  `\u002Frepo-fleet-hygiene:setup apply`, which prefers the portable relative form.\n",{"metadata":24,"description":7,"allowed-tools":28,"argument-hint":30,"user-invocable":31,"disable-model-invocation":20},{"cadence":25,"summary":26,"workflow-stage":27},"weekly","Discover a repository fleet and coordinate read-only evidence handoffs","operator",[29],"Bash(${CLAUDE_SKILL_DIR}\u002Fscripts\u002Faudit-fleet.sh:*)","[[\u003Cdir>]... [--root \u003Cdir>]... [--repo \u003Cdir>]... [--config \u003Cfile>] [--detail]|--apply-plan \u003Cpath>]",true,"main","d0e68c3f36b92a1d6381f0ba0060d7a0b8b886c5","plugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit",7,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins","https:\u002F\u002Fgithub.com\u002Fmelodic-software","https:\u002F\u002Fmelodicsoftware.com","Independent software studio building developer tools, automation, and application services, with roots in music.","2026-11-15 05:15:06",{"source":8,"skillName":5,"commands":42,"prompt":52,"zipAvailable":31},[43,46,49],{"tool":44,"command":45},"npx","npx skills add melodic-software\u002Fclaude-code-plugins --skill audit",{"tool":47,"command":48},"bunx","bunx skills add melodic-software\u002Fclaude-code-plugins --skill audit",{"tool":50,"command":51},"pnpm","pnpm dlx skills add melodic-software\u002Fclaude-code-plugins --skill audit","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002FSKILL.md",[54,60,64,68,72,78],{"path":55,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":58,"sourceHtmlUrl":59},"evals\u002Fevals.json","REFERENCE","参考文档",17489,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002Fevals\u002Fevals.json",{"path":61,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":62,"sourceHtmlUrl":63},"reference\u002Fconfidence-model.md",17086,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002Freference\u002Fconfidence-model.md",{"path":65,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":66,"sourceHtmlUrl":67},"reference\u002Fofficial-sources.md",9117,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002Freference\u002Fofficial-sources.md",{"path":69,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":70,"sourceHtmlUrl":71},"reference\u002Fsecurity-review.md",7973,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002Freference\u002Fsecurity-review.md",{"path":73,"resourceType":74,"resourceTypeDesc":75,"sizeBytes":76,"sourceHtmlUrl":77},"scripts\u002Faudit-fleet.sh","SCRIPT","脚本",172596,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002Fscripts\u002Faudit-fleet.sh",{"path":79,"resourceType":74,"resourceTypeDesc":75,"sizeBytes":80,"sourceHtmlUrl":81},"scripts\u002Faudit-fleet.test.sh",162459,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Frepo-fleet-hygiene\u002Fskills\u002Faudit\u002Fscripts\u002Faudit-fleet.test.sh",[83,90,97,104,109,116],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M7DYYDG31B9YX7N9MB9QW2GW","track","Backlog reads through the bound tracker: search open and closed items, view one item with its comments and state history. Use when another skill searches or reads tracker items.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-flow\u002Fevals\u002Frecall-still-reports-reopened-bug\u002Fwork-items-stub\u002Fskills\u002Ftrack\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":19},"01M7DYXVBAVG3NV30PA7GKW9Q6","reviewer-media","Reviewer media for one pull request: decides whether it needs a narrated walkthrough, builds the storyboard from \u002Freview:explain-change's digest, has \u002Fexplainer-video:produce render it and hands the manifest to \u002Fsource-control:pull-request media. Use when asked for 'reviewer media', 'a walkthrough video for this PR', 'a produced video of this change', or when pull-request ready or monitor hands a PR over.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Freview\u002Fskills\u002Freviewer-media\u002FSKILL.md",[],[],{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":19},"01M7DYRZ5YD44AMYH26YR25GZ8","route","Model and effort per agent role (orchestrator, worker, verifier, retrieval) for a multi-agent run, resolved through the bundled defaults and the user, team and personal override config layers into fenced JSON with single-agent and fan-out variants and the layer behind each value. Use when asked 'which model for the subagents', 'route the roles' or 'model routing', or before a skill passes a role's model on a spawn.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fdiscovery\u002Fevals\u002Fex-model-pin-without-override\u002Fmulti-agent-stub\u002Fskills\u002Froute\u002FSKILL.md",[],[],{"skillNo":105,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":19},"01M7DYRTVT4FWTGDJY0HCCFJMZ","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fdiscovery\u002Fevals\u002Fex-model-override-from-route\u002Fmulti-agent-stub\u002Fskills\u002Froute\u002FSKILL.md",[],[],{"skillNo":110,"name":111,"nameZh":6,"description":112,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":113,"categories":114,"keywords":115,"publishedAt":19},"01M7DYNZC8PNW3T27V2ADKECKJ","principles","Module-design vocabulary for architecture findings, and the checks built on it. Use when: 'what is a deep module', 'what is a seam', 'deep vs shallow module', 'what is the deletion test', 'leverage and locality', 'module design vocabulary', 'which word do we use for X in architecture findings'. Scanning a codebase for friction is \u002Farchitecture:improve.","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Farchitecture\u002Fskills\u002Fprinciples\u002FSKILL.md",[],[],{"skillNo":117,"name":118,"nameZh":6,"description":119,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":120,"categories":121,"keywords":122,"publishedAt":19},"01M7DYNMCZNEWBFHV9KTSS6H60","write","Structured five-field bug report (title, steps to reproduce, expected vs actual, justified severity, suggested fix location) from an informal description, read-only. Use when the user names a defect they observed, or asks for one written up ('report a bug', 'file a bug', 'write this up as a bug').","https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Farchitecture\u002Fevals\u002Fdeepening-defect-goes-to-bugs-write\u002Fbugs-stub\u002Fskills\u002Fwrite\u002FSKILL.md",[],[]]