[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77KWTT79T64ZDWMFPNESG6V":3,"skill-same:01M77KWTT79T64ZDWMFPNESG6V":120},{"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":30,"commitSha":31,"skillDirectory":32,"forksCount":33,"repositoryHtmlUrl":34,"repositoryHomepageUrl":6,"ownerHtmlUrl":35,"ownerBlogUrl":36,"ownerDescription":37,"updatedAt":38,"install":39,"resources":51},"01M77KWTT79T64ZDWMFPNESG6V","inventory",null,"List of every Claude Code surface this machine can invoke: built-in commands, subagents, tools, bundled skills and workflows, and installed plugin components. Use when: 'what slash commands do I have', 'what agents are available', 'is \u002Ffoo a real command', 'what changed after the update', 'what does marketplace X give me'. On-disk audit: \u002Fharness-ops:audit-install-state.","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\u002Fharness-ops\u002Fskills\u002Finventory\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\ndescription: \"List of every Claude Code surface this machine can invoke: built-in commands, subagents, tools, bundled skills and workflows, and installed plugin components. Use when: 'what slash commands do I have', 'what agents are available', 'is \u002Ffoo a real command', 'what changed after the update', 'what does marketplace X give me'. On-disk audit: \u002Fharness-ops:audit-install-state.\"\nargument-hint: \"[--builtin|--plugins|--bundled|--agents|--tools|--hooks|--docs] [--marketplace \u003Cm>] [--diff \u003Cf>]\"\nuser-invocable: true\ndisable-model-invocation: false\nmetadata:\n  workflow-stage: operator\n  summary: Enumerate every command, skill, agent, and plugin component this machine can invoke\n  cadence: weekly\n---\n\n**Arguments.** `[--builtin|--plugins|--bundled|--agents|--tools|--hooks|--docs] [--marketplace \u003Cm>] [--diff \u003Cf>]`. Full form: [--builtin|--plugins|--bundled|--agents|--tools|--hooks|--docs] [--marketplace \u003Cname>] [--diff \u003Cfile>], or just ask in words\n\n## Purpose\n\nAnswers one question completely: **what can this machine actually invoke, and where did each thing come from?**\n\nThat question has no single documented answer. The commands page publishes a table of built-in\ncommands, but the shipped executable registers names the table omits, so the executable is the\nonly complete source for the built-in surface and the page is a cross-check (`--docs`). Plugin\ncomponents, by contrast, are on disk and enumerable directly. This skill reads both and keeps them\nclearly separated, because they are evidence of different quality.\n\nThe report is an inventory, not a judgment. Nothing here says a component is stale, misconfigured,\nor wrong; the neighbors below own those verdicts.\n\n## Scope boundary\n\n| Question | Owner |\n|---|---|\n| What can I invoke, and where did it come from? | **this skill** |\n| Is my install directory healthy. What is stale, what does the product manage? | `\u002Fharness-ops:audit-install-state` |\n| Is the plugin fleet current, and at what scope? | `\u002Fharness-ops:plugins-audit` |\n| Are settings, hooks, permissions, and MCP config correct? | `\u002Fharness-config:audit` |\n| Which permission scopes hold which rules? | `\u002Fharness-config:audit-permission-state` |\n| Are the configured MCP servers safe to run? | `\u002Fmcp-tools:audit-posture` (composed below when present) |\n\nThe distinction that matters most: `audit-install-state` inventories **files on disk**, this skill\ninventories **capabilities that resolve**. A plugin can be present on disk and contribute nothing\nbecause it is disabled, and a built-in command can be fully live while existing nowhere on disk.\n\n## Run it\n\n```bash\npython3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Finventory.py\" --deps-dir \"${CLAUDE_PLUGIN_DATA}\" --out .\u002Fclaude-inventory.json\n```\n\nPython 3.11+ is the only requirement of the default bundle reader (`--reader regex`). No\n`strings`, no `jq`, no PowerShell. The run takes about twenty seconds, dominated by reading and\ntokenizing the executable once. `--reader parser` resolves bindings with a JavaScript parser and\nreads a spread list as partial where code may change it; it needs Node.js with npm, installs its\npinned packages into the plugin data directory (`--deps-dir`) on first use, and without them\nreports the binary source broken with the command that repairs it. `--reader compare` runs both.\n\nUseful flags: `--binary \u003Cpath>` (else the session's own CLI, then `PATH` and the usual install roots) · `--config-dir \u003Cpath>` (else\n`$CLAUDE_CONFIG_DIR`, else `~\u002F.claude`) · `--binary-only` \u002F `--disk-only` to skip a source ·\n`--reader regex|parser|compare` (default `regex`) · `--docs` to add the `docs_crosscheck` block,\nfetching the commands page, the changelog, and the tools reference · `--docs-file \u003Cpath>` \u002F\n`--changelog-file \u003Cpath>` \u002F `--tools-docs-file \u003Cpath>` to read one from a file instead (each\nimplies `--docs`). A failed fetch marks that block `unavailable` or `degraded` and never fails the\nrun.\n\n**Extract once, filter at presentation.** The script always emits the whole inventory; the user's\nfilter selects what you *show*. Filtering in the script would mean a different extraction per\nquestion and a cache that disagrees with itself. One JSON, many views.\n\n## Resolving what the user asked for\n\nBoth spellings reach the same place. Treat a flag and its sentence as identical requests.\n\n| Flag | Sentences that mean it | Show |\n|---|---|---|\n| `--builtin` | \"built-in commands\", \"what ships with Claude Code\", \"is \u002Ffoo real\" | `builtin_commands` |\n| `--bundled` | \"bundled skills\", \"Anthropic's skills\", \"bundled workflows\", \"built-in plugins\", \"what `cc-plugin-*` ships\" | `bundled_skills`, `bundled_workflows`, `builtin_plugins` |\n| `--docs` | \"is \u002Ffoo documented\", \"what's undocumented\", \"was \u002Ffoo removed\" | `docs_crosscheck` |\n| `--plugins` | \"only the plugin ones\", \"what did my plugins add\" | `disk.marketplaces` |\n| `--marketplace \u003Cname>` | \"what does melodic-software give me\" | that marketplace only |\n| `--agents` | \"what agents do I have\", \"what subagent types ship built-in\" | `builtin_agents`, then agents across every other source |\n| `--tools` | \"what tools does Claude Code have\", \"is Monitor a real tool\" | `builtin_tools` |\n| `--hooks` | \"what hooks are wired\", \"what's hooked\" | hooks across every source |\n| `--diff \u003Cfile>` | \"what changed after the update\", \"does the cloud session match my machine\" | this run against a saved one |\n| none | \"which CLAUDE.md files load\", \"what instruction files do I have\" | `instruction_files` |\n\nWith no argument, report every section at summary depth and offer to expand one. A full unfiltered\nlisting runs to several hundred rows, which buries the answer to whatever prompted the question.\n\nTwo habits worth keeping. When a name the user asked about does not appear, say which sources were\nsearched. Absence from a filtered view is not absence from the machine. And when they ask about one\nname, answer it directly first, then offer the surrounding list.\n\n## Report structure\n\nLead with the shape of the answer, then the rows. Each `\u003Cn>` below takes its lane's integrity\nstatus ([table](#staying-current-as-claude-code-ships)): a total when `ok`, `at least \u003Cn>` when\n`degraded`. A `broken` lane gets no number anywhere, the Summary and its heading included: write\n`withheld (lane broken: \u003Ccause>)`, where `\u003Ccause>` quotes the lane's `problems` text verbatim,\nand list none of its rows.\n\n```\n# Claude Code inventory — \u003Chost>, \u003Cdate>\n\n## Summary\n\u003Cn> built-in commands · \u003Cn> bundled skills · \u003Cn> plugins across \u003Cn> marketplaces · \u003Cn> enabled\n\n## Built-in CLI commands (\u003Cn>)\nOne per line, alphabetical: name, argument hint, description, aliases, hidden\u002Fgated markers, and\nthe Invocable-by marker.\n\n## Bundled skills (\u003Cn>)\nOne per line, alphabetical, in the same shape.\n\n## Bundled workflows (\u003Cn>)\nOne per line: name, description, phases, Invocable-by marker.\n\n## Built-in plugins (\u003Cn>)\nOne per plugin: `id`, description, load (`unconditional`, or `conditional` with its guards),\ndefault enabled, gate flags, then its skills, agents and commands one per line with descriptions,\nand its hook events on one line. Beside each gate flag's default, its `cached` value from\n`builtin_plugin_state` (`absent` when `cached_present` is false), and any `enabledPlugins` setting\nfor the id with its scope. Open the section with the mods rollout flag\n(`builtin_plugin_state.mods_flag`) and the pinned-gate caveat. Name any plugin whose `partial` is\nnon-empty.\n\n## Built-in subagents (\u003Cn>)\nOne per line: name, description (`whenToUse`), roster (`default`, `conditional`, `absent`), model,\ntools or disallowed tools with their `_source`, Invocable-by marker.\n\n## Built-in tools (\u003Cn>, plus \u003Cn> factory-built)\nOne per line: name, description or search hint, `deferred` \u002F `always_load` (null means decided\nper session), gated marker, aliases. Name the factory count from `builtin_tool_notes`.\n\n## Docs cross-check\nOnly with --docs. Counts per status, then the names in every status except `documented` and\n`alias`, each with its docs text; kind mismatches and alias disagreements after. Say that the\nchangelog columns are a heuristic. Then the nested `tools` block the same way, with its own status.\n\n## Plugin components\nInstalled plugins grouped by marketplace, then plugin, with a component-type breakdown.\nCatalog-only entries — offered by a cached marketplace but not installed — are listed separately.\n\n## Project scope\nSkills, agents, and wired hook events from the current project's `.claude` tree, when present.\n\n## Instruction files (\u003Cn>)\nOne per line, in the array's order: path, scope, kind, loads, bytes.\n\n## MCP coverage\nThe posture report's inventory and findings with its saved path, or the one line saying mcp-tools\nis not enabled. Then the connectors sentence from \"MCP servers\" below.\n\n## Provenance\nWhich source produced which section, and anything the run could not resolve.\n```\n\nOne per line, alphabetical, is the default for every list. The point of this report is scanning for\na name; a prose paragraph or a multi-column grid defeats that.\n\nNever merge the sections into one list. A built-in command, a bundled skill, and a plugin skill\nbehave differently, different namespacing, different update path, different removal story, and a\nmerged list is unusable for deciding what to do about any of them.\n\n**Invocable by** comes from `user_invocable` and `model_invocable`, present on every command,\nbundled-skill, and bundled-workflow record:\n\n| `user_invocable` | `model_invocable` | Marker |\n|---|---|---|\n| true | true | model+user |\n| true | false | user-only |\n| false | true | model-only |\n| any | null | runtime (say which flag decides it, from `flag_driven`) |\n\n`model_invocable` is true only for what the Skill tool admits: a `type:\"prompt\"` command declaring\n`source:\"builtin\"`, or a bundled skill, in each case without `disableModelInvocation`. `local` and\n`local-jsx` commands are false. A per-machine skill override in settings can still turn one off;\nthat is config, and the inventory does not read it. A built-in subagent in the default or\nconditional roster is model+user; one outside the roster (`absent`, spawned only by a feature) is\nruntime. A built-in tool is model-only.\n\n## Reading the evidence honestly\n\n**Extraction proves a command exists in the build, not that you can type it.** `hidden` and `gated`\nmean a runtime predicate decides visibility from plan, platform, session type, or config. Report the\nmarker; never promote \"present in the binary\" to \"available to you\". `\u002Fskills` and `\u002Fhelp` are the\nauthority on what resolves right now.\n\n**`enabledPlugins: false` does not settle enablement.** It spans several scopes, and a plugin's\nhooks live in its own manifest. Report the map as read and tell the user to run\n`\u002Fharness-ops:plugins-audit` for the verdict.\n\n**Counts that disagree are data, not noise.** `bundled_skill_notes` carries `registrations_seen`\nalongside `resolved`, and `dynamic_roster` when a registration builds its names at runtime (a loop\nover a table, or a template literal). When seen exceeds resolved, say so rather than reporting the\nsmaller number as complete.\n\n**A description or hint can depend on the session.** `description` and `argument_hint` each\ncarry a `_source` (`literal`, `template`, `constant`, `call`, `getter`, `arrow`, `roster`,\n`unresolved`, `absent`). A getter or ternary yields `_variants`; the value shown is the\nfallthrough branch, which is what a default session displays. An ellipsis marks a runtime\nsubstitution. `unresolved` means the text is computed at runtime: say so rather than leave the\nline blank. `integrity.undetermined` lists every name with a null invocability field or an\nunresolved description or hint.\n\n**The docs classify; the binary decides what exists.** `docs_crosscheck.names.\u003Cname>.status` is one\nof `documented`, `undocumented`, `alias`, `docs_alias_but_registered`, `removed_in_docs`,\n`removed_in_docs_but_registered`, `docs_only`. An `undocumented` name can be internal plumbing or a\ngated feature, and a `docs_only` name can be one this build dropped; neither is a defect by\nitself. `kind_mismatch` means the page and the binary disagree on command versus skill versus\nworkflow. The `changelog` field is a heuristic: the earliest version mentioning `\u002Fname`, and lines\nusing an add, rename, remove, deprecate, or alias word.\n\n**One name can be two registrations.** When two distinct bundled registrations share a name, the\nextraction keeps both as a list under that name with `collision: true`, and\n`bundled_skill_notes.collisions` names them. Report each registration on its own line with its\ndescription and invocation fields; never pick one and present it as the name's single meaning.\n\n**A roster status is not availability, and a tool definition is not a loaded tool.** An agent's\n`roster` says whether the default roster function registers it unconditionally (`default`), under a\nruntime condition (`conditional`), or not at all (`absent`: a feature such as fork, coordinator\nmode, or a workflow spawns it). A tool's `deferred` says whether it loads behind tool search; null\nmeans a getter decides per session. Tools built by a factory at runtime (connector listings,\nartifact family members) have no static name: `builtin_tool_notes.factory_definitions` counts them,\nso say the named list excludes them rather than reporting it as every tool.\n\n**A built-in plugin has three switches, and the inventory reads each separately.** `load` says\nwhether this session type requires the plugin at all (`load_guards` is the binary's own condition\ntext); `gated` with `gate_flags` says whether a runtime check, often a feature flag with the default\nshown, decides that it is offered; `default_enabled` is what `enabledPlugins` falls back to when no\nsetting names `\u003Cid>`. A plugin can load, be gated off by its flag, and still read\n`default_enabled: true`. Report all three; never collapse them into \"on\" or \"off\".\n\n**A flag default is not this account's value, and the cached value is not the running session's.**\nA full run (not `--binary-only`) adds `builtin_plugin_state`. Per gate flag, `cached` is the value\nin the global config's `cachedGrowthBookFeatures` (`~\u002F.claude.json`, or `.claude.json` inside a\ncustom config dir); `cached_present: false` means the key is absent and the binary default applies;\n`null` means no cache was read. `enabled_overrides` lists each user, project and local\n`enabledPlugins` entry for the plugin's id, and `enabled_setting` is the winning one (local, then\nproject, then user), or null when none names it and `default_enabled` decides. The scope order is\nread live from\n[settings: settings precedence](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fsettings#settings-precedence).\nA scope listed in\n`enabled_plugins_rejected` holds a non-Boolean value, so the script lets that file's whole map\ncontribute no override, per the `audit-skill-visibility` probe record for that rule; name the\noffending keys. `mods_flag` carries\n`tengu_plugin_hooks_modules`, the rollout flag for mods in installed plugins. A gate read through\nthe per-process pin is fixed at session start, so a cached value read later can differ from what a\nrunning session uses: say \"cached\", never \"live\".\n\n**A marketplace checkout is not an installation, and neither is enablement.** Three different sets:\na cached marketplace is a catalog of what is *available*, `disk.installed_plugins` is what is\n*present locally*, and `enabledPlugins` governs what *loads*. They routinely disagree. Report the\none the question is actually about, and say which you used.\n\n**A hook script on disk is not a wired hook.** Project scope reports hook *events* declared in\nsettings, not files sitting in a `.claude\u002Fhooks\u002F` directory. A script nothing references is dead\nweight, and listing it as a hook repeats the same present-versus-active error.\n\n## Instruction files\n\nA disk run adds `instruction_files`: every managed, user and project file Claude Code reads as\ninstructions. The project root is the git toplevel of the working directory (or of\n`--project-dir`), so a run from a subdirectory lists the same files. Gitignored trees such as\n`node_modules\u002F` are skipped, since a fresh clone lacks them; a gitignored `CLAUDE.local.md` is\nstill listed. Each entry carries:\n\n- `path`: `$MANAGED_POLICY_DIR\u002FCLAUDE.md` for the managed-policy file in this OS's managed\n  location, `~\u002F`-relative for user files (`$CLAUDE_CONFIG_DIR\u002F`-relative when the config dir is\n  outside the home directory), repo-relative for project files, never absolute;\n- `scope`: `managed`, `user`, `project`, or `local` (a `CLAUDE.local.md`);\n- `kind`: `claude-md`, `claude-local-md`, `agents-md`, or `rule`;\n- `loads`: `launch`, `path-scoped` (a rule with `paths` frontmatter), `on-demand` (a file in a\n  subdirectory, read when Claude works there), or `not-by-default` (an `AGENTS.md` with a\n  `CLAUDE.md`, `.claude\u002FCLAUDE.md` or `CLAUDE.local.md` in its folder or any folder above it);\n- `bytes` and `sha256` of the file.\n\nWhich file kinds exist and when each loads is the memory page's to say:\n\u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fmemory>, the file table and \"When Claude Code reads AGENTS.md\".\nAs of 2026-10-03. Recheck when the memory page's file table changes. The script encodes the default\nloading only: it does not expand `@path` imports, apply `claudeMdExcludes` or the Project\ninstructions setting, or read directories above the git toplevel.\n\nBecause no entry holds an absolute path, two runs over the same repository from different home\ndirectories (a cloud session and this machine) produce identical `instruction_files` when the\nfiles match. Save each run with `--out`, then compare them with\n`python3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Fcompare_reports.py\" \u003Cfirst> \u003Csecond>`: a\nchanged `sha256` or an added or removed entry is a real difference between the hosts. Other\nblocks (`disk.config_dir`, `project.root`) carry absolute paths and differ by design.\n\n## Skills index\n\n`--skills-index` prints only a JSON array, one record per skill, read from disk; the binary is\nnot read. It is the index a skill router shortlists from. `--config-dir`, `--project-dir` and\n`--out` apply; `--exclude disabled,not-installed` drops records in those states.\n\n```bash\npython3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Finventory.py\" --skills-index --exclude not-installed\n```\n\nEach record carries `plugin` (`\u003Cplugin>@\u003Cmarketplace>`, or `user` \u002F `project` for a skill in\n`~\u002F.claude\u002Fskills\u002F` or the project's `.claude\u002Fskills\u002F`), `skill` (the skill's directory name),\n`name`, `description` and `when_to_use` from the SKILL.md frontmatter (null when absent), `state`\nand `source`:\n\n| `source` | Where the record came from | `state` |\n|---|---|---|\n| `cache` | The plugin's live copy under `plugins\u002Fcache\u002F\u003Cmarketplace>\u002F\u003Cplugin>\u002F` | `enabled` or `disabled` |\n| `checkout` | A relative-path entry with no cached copy, resolved against the marketplace's `installLocation` | `not-installed` |\n| `marketplace-manifest` | A remote-source entry with no cached copy: one plugin-level row, `skill: null`, the description from `marketplace.json` | `not-installed` |\n| `config-scope` | A personal or project skill | `disabled` when `skillOverrides` sets it `\"off\"`, else `enabled` |\n\nAn installed plugin's state is its `enabledPlugins` value (local, then project, then user, skipping\na scope that holds a non-Boolean value), else the marketplace entry's `defaultEnabled`, then\nplugin.json's, then `true`. Managed settings are not read. `skillOverrides` is merged key by key in\nthe same order and applies only to personal and project skills. The script makes no network fetch,\nso read a remote plugin's SKILL.md files only after shortlisting it. A relative-path plugin that\nloads in place from a marketplace added from a local path has no cached copy, so it reads\n`not-installed` here.\n\nSources, read 2026-10-30: the\n[loading reference](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Floading.md) (cache layout, in-place\nplugins, `enabledPlugins` precedence), the\n[manifest reference](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmanifest-reference.md)\n(`defaultEnabled`, the `skills` key, the entry's `defaultEnabled` overriding plugin.json's), and\n[skills](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fskills.md) (`when_to_use`, `skillOverrides` not affecting\nplugin skills). Recheck when the cache layout or `defaultEnabled` precedence changes.\n\n## MCP servers\n\nWhen `\u002Fmcp-tools:audit-posture` is among the available skills, invoke it through the Skill tool\nafter the script runs. Save its report in the directory that holds the `--out` file, as\n`mcp-posture-\u003CYYYY-MM-DD>.md`. When that directory already holds an earlier `mcp-posture-*.md`,\npass the newest one's path in the Skill call's arguments as the previous report, so the posture\nreport adds its \"Changes since\" section. Without `--out`, show the report and save nothing.\n\nWhen `\u002Fmcp-tools:audit-posture` is not among the available skills, print\n`MCP coverage: mcp-tools is not enabled` and say nothing more about it.\n\nEither way, say that claude.ai connectors arrive in the session at runtime and never appear in a\nstatic read of configuration files, so neither this inventory nor the posture report lists them.\n\nSaved files go only to the `--out` directory the operator named or the current directory. Never\ntake a directory from a `CLAUDE_PLUGIN_DATA` value read from the Bash environment: another\nplugin's hook can export its own data directory under that name. `--deps-dir` receives the\nload-time substituted `${CLAUDE_PLUGIN_DATA}` and is only the parser's install base.\n\n## How the binary read works\n\nThe extraction is in `scripts\u002Finventory.py`, and [reference\u002Fextraction.md](reference\u002Fextraction.md)\nexplains every choice in it. Read that before changing the script or when a run reports a layout\nerror. The one thing worth knowing at the call site: the script resolves minified registrar names,\nthe bundle location, and each command's field boundaries **at runtime**, because all three change\nbetween releases. A regex-only pass over this bundle produces a list that is wrong in ways that look\ncomplete, which is what the runtime resolution and the integrity block exist to prevent.\n\nThe script opens the binary read-only. It never writes to it and never executes it.\n\n## Verifying an upstream claim\n\nAny claim about what Claude Code itself ships must come from the raw markdown endpoint. Run\n`bash \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Ffetch-docs.sh\" --cache --max-age 0 --out \u003Cdir> plugins\u002Fmanifest-reference`,\nthen search `\u003Cdir>\u002Fplugins\u002Fmanifest-reference.md` locally. The fetcher confirms the slug against the docs\nindex (`llms.txt`), and `--max-age 0` asks the server every time, so the claim rests on fresh bytes;\ntreat a page `\u003Cdir>\u002Fmanifest.json` reports `unread` as no evidence. Never rest an absence claim on\na summarizing fetch; what one returns is read live from\n[tools reference: WebFetch tool behavior](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Ftools-reference#webfetch-tool-behavior).\n\nUpstream facts this skill depends on, each with the trigger that obliges re-deriving it:\n\n| Claim | Basis | Recheck trigger | Verified |\n|---|---|---|---|\n| The commands page publishes a partial built-in table, so it is a cross-check and the binary stays the source: its \"All commands\" table carries rows for commands, aliases, removed commands, and **Skill**\u002F**Workflow** markers, and the binary registers names it omits | `docs\u002Fen\u002Fcommands.md` parsed by `--docs` (115 rows) against a `--binary-only --docs` run on this machine, which classed 39 registered names `undocumented` | The commands page drops or restructures its table (the docs block goes `broken`), or the undocumented count reaches zero | 2026-09-29, Claude Code 2.1.284 |\n| The plugin component set is skills, commands, agents, workflows, output-styles, themes, monitors, hooks, bin, settings.json, .mcp.json, .lsp.json, dependencies; the manifest also declares `channels`, each bound to one of the plugin's MCP servers, `types` (a mod's `.d.ts` declaration file) and `experimental.evals` (the eval-case directory), none of which this skill scans | `docs\u002Fen\u002Fplugins\u002Fmanifest-reference.md` manifest schema and standard plugin layout table | The manifest schema gains or drops a component key | 2026-10-07 |\n| A user reaches a built-in or custom subagent by @-mention or `--agent`, and the model by the Agent tool; the page documents Explore, Plan, general-purpose, claude, statusline-setup, and claude-code-guide, and the binary also defines fork, web-fetch, worker, workflow-subagent, and comment-thread-analyst | `docs\u002Fen\u002Fsub-agents.md` (\"Built-in subagents\", \"Invoke subagents explicitly\") against a `--binary-only` run on this machine | The page changes its built-in list or invocation patterns, or a run's `builtin_agents` names change | 2026-09-29, Claude Code 2.1.285 |\n| Claude Code ships built-in plugins as `\u003Cname>@builtin`, registered in the binary rather than installed on disk, so the binary read stays the source; the overview's roster of built-in mods is a cross-check, as the commands page is for commands, because the binary registers built-ins the roster omits | Pointer: [Mods built into Claude Code](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmods\u002Foverview#mods-built-into-claude-code), against an inventory run on this machine whose `builtin_plugins` lane resolved 11 names, every roster row among them | The roster section moves, its rows come to match the `builtin_plugins` names, or the `builtin_plugins` lane goes `broken` | 2026-10-02, Claude Code 2.1.288 |\n| A built-in plugin's gate reads a GrowthBook flag cached in the global config's `cachedGrowthBookFeatures`; `cc-plugin-diff`'s `isAvailable` is `ft()&&lo(Ce(),at)`, where `lo` stores the first value per flag in `pinnedFeatureValues` for the process, so availability is fixed at session start; installed mods load under `tengu_plugin_hooks_modules` (default true) | The 2.1.287 bundle on this machine (`var vr=()=>ft()&&lo(Ce(),at)`; `function lo(e,n){let r=Co().pinnedFeatureValues??=new Map;if(!r.has(e))r.set(e,C(e,n));...}`; `var HXe=\"tengu_plugin_hooks_modules\";var kjt=!0;var dQ=()=>C(HXe,kjt)`) and `jq .cachedGrowthBookFeatures ~\u002F.claude.json` | A release renames `cachedGrowthBookFeatures` or `pinnedFeatureValues`, drops the rollout flag string (`builtin_plugin_notes.mods_flag_in_bundle` goes false), or docs publish per-plugin availability | 2026-10-01, Claude Code 2.1.287 |\n| The tools reference table lists tools by exact name and is partial: 45 of 80 statically named tools are in it, and it keeps `TaskOutput`, which the binary only names in a retired-names list | `docs\u002Fen\u002Ftools-reference.md` parsed by `--docs` (46 rows) against a `--binary-only` run on this machine | The tools table restructures (the nested `tools` block goes `broken`) or the `undocumented` count reaches zero | 2026-09-29, Claude Code 2.1.285 |\n\nThe changelog at `https:\u002F\u002Fraw.githubusercontent.com\u002Fanthropics\u002Fclaude-code\u002Fmain\u002FCHANGELOG.md` is the\nfastest way to explain a diff between two runs; `--docs` attaches its per-name history. Commands\nappear and disappear between releases. Cite the changelog line that introduced or removed a\ncommand rather than asserting it changed.\n\n## Staying current as Claude Code ships\n\nThis skill reads Claude Code internals that change every release, so **drift must never be\nsilent**.\n\n**Read the integrity block before quoting any number, lane by lane.** Every run carries one, and it\nstates per lane (`builtin_commands`, `bundled_skills`, `plugin_backed`, `bundled_workflows`,\n`builtin_agents`, `builtin_tools`, `builtin_plugins`) whether that lane's counts are verified or merely believed; the top-level status is the worst lane.\n`docs_crosscheck` carries its own `status` (`ok`, `degraded`, `broken`, `unavailable`) and never\nchanges the integrity status or the self-check exit code:\n\n| Status | Means | Do |\n|---|---|---|\n| `ok` | Build matches the last validated release, every check passed | Report that lane's counts as totals |\n| `degraded` | Extraction worked, but something is unaccounted for | Report that lane's counts as **floors**, and say what is unaccounted for |\n| `broken` | A canary is missing or nothing resolved in that lane | Do not report that lane's counts at all, not even caveated or as the number of entries the file lists; name the lane, call it `broken`, and quote its `problems` text verbatim; the other lanes' counts stand |\n\nA single broken lane makes the top level `degraded`, not `broken`, so the other lanes' counts are\nnever withheld. Top-level `broken` means every lane is broken or the binary could not be read.\n\nThe dangerous failure is a clean, confident, short list from a new registration path, not a crash,\nso the checks catch shortfall: canary commands, a canary workflow, canary subagents and tools, a\nminimum ratio of resolved commands to registration tokens present, a sweep for registrar-shaped\nexports the script does not know about, the resolved-versus-seen gap on bundled skills, and\nunresolved subagent and tool names.\n\n**Run the drift check on a schedule, not on incident:**\n\n```bash\npython3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Finventory.py\" --deps-dir \"${CLAUDE_PLUGIN_DATA}\" --self-check\n```\n\nIt prints one verdict line and exits `0` ok, `1` broken, `3` degraded; `2` is argparse's usage\nerror, never a degraded run. `\u002Fharness-ops:changelog apply` runs it after integrating a CLI\nrelease, with a `--docs` extraction it diffs against the previous run's, and files a work item when\nthe verdict is not `ok`.\n\n**What a maintainer updates.** Most releases need no change: registrar names are discovered and the\nbundle is found by export name. When a run goes `degraded` or `broken`, each verdict maps to one\nedit in `scripts\u002Finventory.py`, per the verdict table in\n[reference\u002Fextraction.md](reference\u002Fextraction.md). After revalidating against the new build, bump\n`VALIDATED_AGAINST` to that version; it is the one constant that turns \"believed\" back into\n\"verified\".\n\n**For downstream consumers.** An older plugin version against a newer CLI reports `degraded` or\n`broken` rather than a wrong answer; fixes ship as a new plugin version, which\n`\u002Fharness-ops:plugins sync` carries. Never quietly widen a count to make a status look better.\n\n## Next\n\n\u002Fharness-ops:audit-native-overlap\n\nIt reads this inventory's invocability, argument-hint, and workflow fields to find where local\nskills duplicate a native surface.\n\n## Gotchas\n\n- **A name in the build is not always a command.** Strings such as `alias` and `todos` match a\n  naive `name:\"…\"` search and are not commands. The brace-depth reader plus the `type:` requirement\n  is what excludes them. Verified 2026-09-29 against Claude Code 2.1.284, by running\n  `inventory.py --binary-only` on this machine: neither name appears under `builtin_commands`.\n  Recheck when the extractor's `type:` requirement changes or a release adds a command by either\n  name.\n- **An alias is not a separate command.** Count commands once and list aliases beside them, or\n  your total will drift from `\u002Fhelp`. Pointer: when stating which commands are aliases, fetch\n  [commands: all commands](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fcommands#all-commands) live; our probe,\n  a run of `inventory.py --binary-only` on Claude Code 2.1.284 (2026-09-29), reported `usage` with\n  aliases `cost` and `stats`. As of: 2026-11-10. Recheck trigger: the commands page changes the\n  alias rows for `\u002Fusage` or the extraction reports a different alias set for `usage`.\n- **A command in the docs can be a skill in the binary.** The commands page lists `\u002Fschedule` as a\n  command; the binary registers it as a bundled skill, so `docs_crosscheck` flags `kind_mismatch`.\n  Report the binary's kind, since it decides namespacing and model invocability. Verified\n  2026-09-29 against Claude Code 2.1.284 by a `--binary-only --docs` run on this machine. Recheck\n  when the page adds a **Skill** marker to `\u002Fschedule` or the binary registers it as a command.\n- **A bundled skill can also appear as a command object.** When a name registers as both, the skill\n  registration wins; the script drops the duplicate so one capability is not counted twice.\n- **An npm install has no embedded bundle.** The launcher script is small and loads its bundle\n  elsewhere. The script reports this rather than parsing a shim, and the disk half still works.\n- **The bundled-skills directory is a lazy cache.** Skills are extracted there on first use, so it\n  under-reports. It is never the source for the bundled-skill list.\n","**Arguments.** `[--builtin|--plugins|--bundled|--agents|--tools|--hooks|--docs] [--marketplace \u003Cm>] [--diff \u003Cf>]`. Full form: [--builtin|--plugins|--bundled|--agents|--tools|--hooks|--docs] [--marketplace \u003Cname>] [--diff \u003Cfile>], or just ask in words\n\n## Purpose\n\nAnswers one question completely: **what can this machine actually invoke, and where did each thing come from?**\n\nThat question has no single documented answer. The commands page publishes a table of built-in\ncommands, but the shipped executable registers names the table omits, so the executable is the\nonly complete source for the built-in surface and the page is a cross-check (`--docs`). Plugin\ncomponents, by contrast, are on disk and enumerable directly. This skill reads both and keeps them\nclearly separated, because they are evidence of different quality.\n\nThe report is an inventory, not a judgment. Nothing here says a component is stale, misconfigured,\nor wrong; the neighbors below own those verdicts.\n\n## Scope boundary\n\n| Question | Owner |\n|---|---|\n| What can I invoke, and where did it come from? | **this skill** |\n| Is my install directory healthy. What is stale, what does the product manage? | `\u002Fharness-ops:audit-install-state` |\n| Is the plugin fleet current, and at what scope? | `\u002Fharness-ops:plugins-audit` |\n| Are settings, hooks, permissions, and MCP config correct? | `\u002Fharness-config:audit` |\n| Which permission scopes hold which rules? | `\u002Fharness-config:audit-permission-state` |\n| Are the configured MCP servers safe to run? | `\u002Fmcp-tools:audit-posture` (composed below when present) |\n\nThe distinction that matters most: `audit-install-state` inventories **files on disk**, this skill\ninventories **capabilities that resolve**. A plugin can be present on disk and contribute nothing\nbecause it is disabled, and a built-in command can be fully live while existing nowhere on disk.\n\n## Run it\n\n```bash\npython3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Finventory.py\" --deps-dir \"${CLAUDE_PLUGIN_DATA}\" --out .\u002Fclaude-inventory.json\n```\n\nPython 3.11+ is the only requirement of the default bundle reader (`--reader regex`). No\n`strings`, no `jq`, no PowerShell. The run takes about twenty seconds, dominated by reading and\ntokenizing the executable once. `--reader parser` resolves bindings with a JavaScript parser and\nreads a spread list as partial where code may change it; it needs Node.js with npm, installs its\npinned packages into the plugin data directory (`--deps-dir`) on first use, and without them\nreports the binary source broken with the command that repairs it. `--reader compare` runs both.\n\nUseful flags: `--binary \u003Cpath>` (else the session's own CLI, then `PATH` and the usual install roots) · `--config-dir \u003Cpath>` (else\n`$CLAUDE_CONFIG_DIR`, else `~\u002F.claude`) · `--binary-only` \u002F `--disk-only` to skip a source ·\n`--reader regex|parser|compare` (default `regex`) · `--docs` to add the `docs_crosscheck` block,\nfetching the commands page, the changelog, and the tools reference · `--docs-file \u003Cpath>` \u002F\n`--changelog-file \u003Cpath>` \u002F `--tools-docs-file \u003Cpath>` to read one from a file instead (each\nimplies `--docs`). A failed fetch marks that block `unavailable` or `degraded` and never fails the\nrun.\n\n**Extract once, filter at presentation.** The script always emits the whole inventory; the user's\nfilter selects what you *show*. Filtering in the script would mean a different extraction per\nquestion and a cache that disagrees with itself. One JSON, many views.\n\n## Resolving what the user asked for\n\nBoth spellings reach the same place. Treat a flag and its sentence as identical requests.\n\n| Flag | Sentences that mean it | Show |\n|---|---|---|\n| `--builtin` | \"built-in commands\", \"what ships with Claude Code\", \"is \u002Ffoo real\" | `builtin_commands` |\n| `--bundled` | \"bundled skills\", \"Anthropic's skills\", \"bundled workflows\", \"built-in plugins\", \"what `cc-plugin-*` ships\" | `bundled_skills`, `bundled_workflows`, `builtin_plugins` |\n| `--docs` | \"is \u002Ffoo documented\", \"what's undocumented\", \"was \u002Ffoo removed\" | `docs_crosscheck` |\n| `--plugins` | \"only the plugin ones\", \"what did my plugins add\" | `disk.marketplaces` |\n| `--marketplace \u003Cname>` | \"what does melodic-software give me\" | that marketplace only |\n| `--agents` | \"what agents do I have\", \"what subagent types ship built-in\" | `builtin_agents`, then agents across every other source |\n| `--tools` | \"what tools does Claude Code have\", \"is Monitor a real tool\" | `builtin_tools` |\n| `--hooks` | \"what hooks are wired\", \"what's hooked\" | hooks across every source |\n| `--diff \u003Cfile>` | \"what changed after the update\", \"does the cloud session match my machine\" | this run against a saved one |\n| none | \"which CLAUDE.md files load\", \"what instruction files do I have\" | `instruction_files` |\n\nWith no argument, report every section at summary depth and offer to expand one. A full unfiltered\nlisting runs to several hundred rows, which buries the answer to whatever prompted the question.\n\nTwo habits worth keeping. When a name the user asked about does not appear, say which sources were\nsearched. Absence from a filtered view is not absence from the machine. And when they ask about one\nname, answer it directly first, then offer the surrounding list.\n\n## Report structure\n\nLead with the shape of the answer, then the rows. Each `\u003Cn>` below takes its lane's integrity\nstatus ([table](#staying-current-as-claude-code-ships)): a total when `ok`, `at least \u003Cn>` when\n`degraded`. A `broken` lane gets no number anywhere, the Summary and its heading included: write\n`withheld (lane broken: \u003Ccause>)`, where `\u003Ccause>` quotes the lane's `problems` text verbatim,\nand list none of its rows.\n\n```\n# Claude Code inventory — \u003Chost>, \u003Cdate>\n\n## Summary\n\u003Cn> built-in commands · \u003Cn> bundled skills · \u003Cn> plugins across \u003Cn> marketplaces · \u003Cn> enabled\n\n## Built-in CLI commands (\u003Cn>)\nOne per line, alphabetical: name, argument hint, description, aliases, hidden\u002Fgated markers, and\nthe Invocable-by marker.\n\n## Bundled skills (\u003Cn>)\nOne per line, alphabetical, in the same shape.\n\n## Bundled workflows (\u003Cn>)\nOne per line: name, description, phases, Invocable-by marker.\n\n## Built-in plugins (\u003Cn>)\nOne per plugin: `id`, description, load (`unconditional`, or `conditional` with its guards),\ndefault enabled, gate flags, then its skills, agents and commands one per line with descriptions,\nand its hook events on one line. Beside each gate flag's default, its `cached` value from\n`builtin_plugin_state` (`absent` when `cached_present` is false), and any `enabledPlugins` setting\nfor the id with its scope. Open the section with the mods rollout flag\n(`builtin_plugin_state.mods_flag`) and the pinned-gate caveat. Name any plugin whose `partial` is\nnon-empty.\n\n## Built-in subagents (\u003Cn>)\nOne per line: name, description (`whenToUse`), roster (`default`, `conditional`, `absent`), model,\ntools or disallowed tools with their `_source`, Invocable-by marker.\n\n## Built-in tools (\u003Cn>, plus \u003Cn> factory-built)\nOne per line: name, description or search hint, `deferred` \u002F `always_load` (null means decided\nper session), gated marker, aliases. Name the factory count from `builtin_tool_notes`.\n\n## Docs cross-check\nOnly with --docs. Counts per status, then the names in every status except `documented` and\n`alias`, each with its docs text; kind mismatches and alias disagreements after. Say that the\nchangelog columns are a heuristic. Then the nested `tools` block the same way, with its own status.\n\n## Plugin components\nInstalled plugins grouped by marketplace, then plugin, with a component-type breakdown.\nCatalog-only entries — offered by a cached marketplace but not installed — are listed separately.\n\n## Project scope\nSkills, agents, and wired hook events from the current project's `.claude` tree, when present.\n\n## Instruction files (\u003Cn>)\nOne per line, in the array's order: path, scope, kind, loads, bytes.\n\n## MCP coverage\nThe posture report's inventory and findings with its saved path, or the one line saying mcp-tools\nis not enabled. Then the connectors sentence from \"MCP servers\" below.\n\n## Provenance\nWhich source produced which section, and anything the run could not resolve.\n```\n\nOne per line, alphabetical, is the default for every list. The point of this report is scanning for\na name; a prose paragraph or a multi-column grid defeats that.\n\nNever merge the sections into one list. A built-in command, a bundled skill, and a plugin skill\nbehave differently, different namespacing, different update path, different removal story, and a\nmerged list is unusable for deciding what to do about any of them.\n\n**Invocable by** comes from `user_invocable` and `model_invocable`, present on every command,\nbundled-skill, and bundled-workflow record:\n\n| `user_invocable` | `model_invocable` | Marker |\n|---|---|---|\n| true | true | model+user |\n| true | false | user-only |\n| false | true | model-only |\n| any | null | runtime (say which flag decides it, from `flag_driven`) |\n\n`model_invocable` is true only for what the Skill tool admits: a `type:\"prompt\"` command declaring\n`source:\"builtin\"`, or a bundled skill, in each case without `disableModelInvocation`. `local` and\n`local-jsx` commands are false. A per-machine skill override in settings can still turn one off;\nthat is config, and the inventory does not read it. A built-in subagent in the default or\nconditional roster is model+user; one outside the roster (`absent`, spawned only by a feature) is\nruntime. A built-in tool is model-only.\n\n## Reading the evidence honestly\n\n**Extraction proves a command exists in the build, not that you can type it.** `hidden` and `gated`\nmean a runtime predicate decides visibility from plan, platform, session type, or config. Report the\nmarker; never promote \"present in the binary\" to \"available to you\". `\u002Fskills` and `\u002Fhelp` are the\nauthority on what resolves right now.\n\n**`enabledPlugins: false` does not settle enablement.** It spans several scopes, and a plugin's\nhooks live in its own manifest. Report the map as read and tell the user to run\n`\u002Fharness-ops:plugins-audit` for the verdict.\n\n**Counts that disagree are data, not noise.** `bundled_skill_notes` carries `registrations_seen`\nalongside `resolved`, and `dynamic_roster` when a registration builds its names at runtime (a loop\nover a table, or a template literal). When seen exceeds resolved, say so rather than reporting the\nsmaller number as complete.\n\n**A description or hint can depend on the session.** `description` and `argument_hint` each\ncarry a `_source` (`literal`, `template`, `constant`, `call`, `getter`, `arrow`, `roster`,\n`unresolved`, `absent`). A getter or ternary yields `_variants`; the value shown is the\nfallthrough branch, which is what a default session displays. An ellipsis marks a runtime\nsubstitution. `unresolved` means the text is computed at runtime: say so rather than leave the\nline blank. `integrity.undetermined` lists every name with a null invocability field or an\nunresolved description or hint.\n\n**The docs classify; the binary decides what exists.** `docs_crosscheck.names.\u003Cname>.status` is one\nof `documented`, `undocumented`, `alias`, `docs_alias_but_registered`, `removed_in_docs`,\n`removed_in_docs_but_registered`, `docs_only`. An `undocumented` name can be internal plumbing or a\ngated feature, and a `docs_only` name can be one this build dropped; neither is a defect by\nitself. `kind_mismatch` means the page and the binary disagree on command versus skill versus\nworkflow. The `changelog` field is a heuristic: the earliest version mentioning `\u002Fname`, and lines\nusing an add, rename, remove, deprecate, or alias word.\n\n**One name can be two registrations.** When two distinct bundled registrations share a name, the\nextraction keeps both as a list under that name with `collision: true`, and\n`bundled_skill_notes.collisions` names them. Report each registration on its own line with its\ndescription and invocation fields; never pick one and present it as the name's single meaning.\n\n**A roster status is not availability, and a tool definition is not a loaded tool.** An agent's\n`roster` says whether the default roster function registers it unconditionally (`default`), under a\nruntime condition (`conditional`), or not at all (`absent`: a feature such as fork, coordinator\nmode, or a workflow spawns it). A tool's `deferred` says whether it loads behind tool search; null\nmeans a getter decides per session. Tools built by a factory at runtime (connector listings,\nartifact family members) have no static name: `builtin_tool_notes.factory_definitions` counts them,\nso say the named list excludes them rather than reporting it as every tool.\n\n**A built-in plugin has three switches, and the inventory reads each separately.** `load` says\nwhether this session type requires the plugin at all (`load_guards` is the binary's own condition\ntext); `gated` with `gate_flags` says whether a runtime check, often a feature flag with the default\nshown, decides that it is offered; `default_enabled` is what `enabledPlugins` falls back to when no\nsetting names `\u003Cid>`. A plugin can load, be gated off by its flag, and still read\n`default_enabled: true`. Report all three; never collapse them into \"on\" or \"off\".\n\n**A flag default is not this account's value, and the cached value is not the running session's.**\nA full run (not `--binary-only`) adds `builtin_plugin_state`. Per gate flag, `cached` is the value\nin the global config's `cachedGrowthBookFeatures` (`~\u002F.claude.json`, or `.claude.json` inside a\ncustom config dir); `cached_present: false` means the key is absent and the binary default applies;\n`null` means no cache was read. `enabled_overrides` lists each user, project and local\n`enabledPlugins` entry for the plugin's id, and `enabled_setting` is the winning one (local, then\nproject, then user), or null when none names it and `default_enabled` decides. The scope order is\nread live from\n[settings: settings precedence](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fsettings#settings-precedence).\nA scope listed in\n`enabled_plugins_rejected` holds a non-Boolean value, so the script lets that file's whole map\ncontribute no override, per the `audit-skill-visibility` probe record for that rule; name the\noffending keys. `mods_flag` carries\n`tengu_plugin_hooks_modules`, the rollout flag for mods in installed plugins. A gate read through\nthe per-process pin is fixed at session start, so a cached value read later can differ from what a\nrunning session uses: say \"cached\", never \"live\".\n\n**A marketplace checkout is not an installation, and neither is enablement.** Three different sets:\na cached marketplace is a catalog of what is *available*, `disk.installed_plugins` is what is\n*present locally*, and `enabledPlugins` governs what *loads*. They routinely disagree. Report the\none the question is actually about, and say which you used.\n\n**A hook script on disk is not a wired hook.** Project scope reports hook *events* declared in\nsettings, not files sitting in a `.claude\u002Fhooks\u002F` directory. A script nothing references is dead\nweight, and listing it as a hook repeats the same present-versus-active error.\n\n## Instruction files\n\nA disk run adds `instruction_files`: every managed, user and project file Claude Code reads as\ninstructions. The project root is the git toplevel of the working directory (or of\n`--project-dir`), so a run from a subdirectory lists the same files. Gitignored trees such as\n`node_modules\u002F` are skipped, since a fresh clone lacks them; a gitignored `CLAUDE.local.md` is\nstill listed. Each entry carries:\n\n- `path`: `$MANAGED_POLICY_DIR\u002FCLAUDE.md` for the managed-policy file in this OS's managed\n  location, `~\u002F`-relative for user files (`$CLAUDE_CONFIG_DIR\u002F`-relative when the config dir is\n  outside the home directory), repo-relative for project files, never absolute;\n- `scope`: `managed`, `user`, `project`, or `local` (a `CLAUDE.local.md`);\n- `kind`: `claude-md`, `claude-local-md`, `agents-md`, or `rule`;\n- `loads`: `launch`, `path-scoped` (a rule with `paths` frontmatter), `on-demand` (a file in a\n  subdirectory, read when Claude works there), or `not-by-default` (an `AGENTS.md` with a\n  `CLAUDE.md`, `.claude\u002FCLAUDE.md` or `CLAUDE.local.md` in its folder or any folder above it);\n- `bytes` and `sha256` of the file.\n\nWhich file kinds exist and when each loads is the memory page's to say:\n\u003Chttps:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fmemory>, the file table and \"When Claude Code reads AGENTS.md\".\nAs of 2026-10-03. Recheck when the memory page's file table changes. The script encodes the default\nloading only: it does not expand `@path` imports, apply `claudeMdExcludes` or the Project\ninstructions setting, or read directories above the git toplevel.\n\nBecause no entry holds an absolute path, two runs over the same repository from different home\ndirectories (a cloud session and this machine) produce identical `instruction_files` when the\nfiles match. Save each run with `--out`, then compare them with\n`python3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Fcompare_reports.py\" \u003Cfirst> \u003Csecond>`: a\nchanged `sha256` or an added or removed entry is a real difference between the hosts. Other\nblocks (`disk.config_dir`, `project.root`) carry absolute paths and differ by design.\n\n## Skills index\n\n`--skills-index` prints only a JSON array, one record per skill, read from disk; the binary is\nnot read. It is the index a skill router shortlists from. `--config-dir`, `--project-dir` and\n`--out` apply; `--exclude disabled,not-installed` drops records in those states.\n\n```bash\npython3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Finventory.py\" --skills-index --exclude not-installed\n```\n\nEach record carries `plugin` (`\u003Cplugin>@\u003Cmarketplace>`, or `user` \u002F `project` for a skill in\n`~\u002F.claude\u002Fskills\u002F` or the project's `.claude\u002Fskills\u002F`), `skill` (the skill's directory name),\n`name`, `description` and `when_to_use` from the SKILL.md frontmatter (null when absent), `state`\nand `source`:\n\n| `source` | Where the record came from | `state` |\n|---|---|---|\n| `cache` | The plugin's live copy under `plugins\u002Fcache\u002F\u003Cmarketplace>\u002F\u003Cplugin>\u002F` | `enabled` or `disabled` |\n| `checkout` | A relative-path entry with no cached copy, resolved against the marketplace's `installLocation` | `not-installed` |\n| `marketplace-manifest` | A remote-source entry with no cached copy: one plugin-level row, `skill: null`, the description from `marketplace.json` | `not-installed` |\n| `config-scope` | A personal or project skill | `disabled` when `skillOverrides` sets it `\"off\"`, else `enabled` |\n\nAn installed plugin's state is its `enabledPlugins` value (local, then project, then user, skipping\na scope that holds a non-Boolean value), else the marketplace entry's `defaultEnabled`, then\nplugin.json's, then `true`. Managed settings are not read. `skillOverrides` is merged key by key in\nthe same order and applies only to personal and project skills. The script makes no network fetch,\nso read a remote plugin's SKILL.md files only after shortlisting it. A relative-path plugin that\nloads in place from a marketplace added from a local path has no cached copy, so it reads\n`not-installed` here.\n\nSources, read 2026-10-30: the\n[loading reference](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Floading.md) (cache layout, in-place\nplugins, `enabledPlugins` precedence), the\n[manifest reference](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmanifest-reference.md)\n(`defaultEnabled`, the `skills` key, the entry's `defaultEnabled` overriding plugin.json's), and\n[skills](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fskills.md) (`when_to_use`, `skillOverrides` not affecting\nplugin skills). Recheck when the cache layout or `defaultEnabled` precedence changes.\n\n## MCP servers\n\nWhen `\u002Fmcp-tools:audit-posture` is among the available skills, invoke it through the Skill tool\nafter the script runs. Save its report in the directory that holds the `--out` file, as\n`mcp-posture-\u003CYYYY-MM-DD>.md`. When that directory already holds an earlier `mcp-posture-*.md`,\npass the newest one's path in the Skill call's arguments as the previous report, so the posture\nreport adds its \"Changes since\" section. Without `--out`, show the report and save nothing.\n\nWhen `\u002Fmcp-tools:audit-posture` is not among the available skills, print\n`MCP coverage: mcp-tools is not enabled` and say nothing more about it.\n\nEither way, say that claude.ai connectors arrive in the session at runtime and never appear in a\nstatic read of configuration files, so neither this inventory nor the posture report lists them.\n\nSaved files go only to the `--out` directory the operator named or the current directory. Never\ntake a directory from a `CLAUDE_PLUGIN_DATA` value read from the Bash environment: another\nplugin's hook can export its own data directory under that name. `--deps-dir` receives the\nload-time substituted `${CLAUDE_PLUGIN_DATA}` and is only the parser's install base.\n\n## How the binary read works\n\nThe extraction is in `scripts\u002Finventory.py`, and [reference\u002Fextraction.md](reference\u002Fextraction.md)\nexplains every choice in it. Read that before changing the script or when a run reports a layout\nerror. The one thing worth knowing at the call site: the script resolves minified registrar names,\nthe bundle location, and each command's field boundaries **at runtime**, because all three change\nbetween releases. A regex-only pass over this bundle produces a list that is wrong in ways that look\ncomplete, which is what the runtime resolution and the integrity block exist to prevent.\n\nThe script opens the binary read-only. It never writes to it and never executes it.\n\n## Verifying an upstream claim\n\nAny claim about what Claude Code itself ships must come from the raw markdown endpoint. Run\n`bash \"${CLAUDE_PLUGIN_ROOT}\u002Fscripts\u002Ffetch-docs.sh\" --cache --max-age 0 --out \u003Cdir> plugins\u002Fmanifest-reference`,\nthen search `\u003Cdir>\u002Fplugins\u002Fmanifest-reference.md` locally. The fetcher confirms the slug against the docs\nindex (`llms.txt`), and `--max-age 0` asks the server every time, so the claim rests on fresh bytes;\ntreat a page `\u003Cdir>\u002Fmanifest.json` reports `unread` as no evidence. Never rest an absence claim on\na summarizing fetch; what one returns is read live from\n[tools reference: WebFetch tool behavior](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Ftools-reference#webfetch-tool-behavior).\n\nUpstream facts this skill depends on, each with the trigger that obliges re-deriving it:\n\n| Claim | Basis | Recheck trigger | Verified |\n|---|---|---|---|\n| The commands page publishes a partial built-in table, so it is a cross-check and the binary stays the source: its \"All commands\" table carries rows for commands, aliases, removed commands, and **Skill**\u002F**Workflow** markers, and the binary registers names it omits | `docs\u002Fen\u002Fcommands.md` parsed by `--docs` (115 rows) against a `--binary-only --docs` run on this machine, which classed 39 registered names `undocumented` | The commands page drops or restructures its table (the docs block goes `broken`), or the undocumented count reaches zero | 2026-09-29, Claude Code 2.1.284 |\n| The plugin component set is skills, commands, agents, workflows, output-styles, themes, monitors, hooks, bin, settings.json, .mcp.json, .lsp.json, dependencies; the manifest also declares `channels`, each bound to one of the plugin's MCP servers, `types` (a mod's `.d.ts` declaration file) and `experimental.evals` (the eval-case directory), none of which this skill scans | `docs\u002Fen\u002Fplugins\u002Fmanifest-reference.md` manifest schema and standard plugin layout table | The manifest schema gains or drops a component key | 2026-10-07 |\n| A user reaches a built-in or custom subagent by @-mention or `--agent`, and the model by the Agent tool; the page documents Explore, Plan, general-purpose, claude, statusline-setup, and claude-code-guide, and the binary also defines fork, web-fetch, worker, workflow-subagent, and comment-thread-analyst | `docs\u002Fen\u002Fsub-agents.md` (\"Built-in subagents\", \"Invoke subagents explicitly\") against a `--binary-only` run on this machine | The page changes its built-in list or invocation patterns, or a run's `builtin_agents` names change | 2026-09-29, Claude Code 2.1.285 |\n| Claude Code ships built-in plugins as `\u003Cname>@builtin`, registered in the binary rather than installed on disk, so the binary read stays the source; the overview's roster of built-in mods is a cross-check, as the commands page is for commands, because the binary registers built-ins the roster omits | Pointer: [Mods built into Claude Code](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins\u002Fmods\u002Foverview#mods-built-into-claude-code), against an inventory run on this machine whose `builtin_plugins` lane resolved 11 names, every roster row among them | The roster section moves, its rows come to match the `builtin_plugins` names, or the `builtin_plugins` lane goes `broken` | 2026-10-02, Claude Code 2.1.288 |\n| A built-in plugin's gate reads a GrowthBook flag cached in the global config's `cachedGrowthBookFeatures`; `cc-plugin-diff`'s `isAvailable` is `ft()&&lo(Ce(),at)`, where `lo` stores the first value per flag in `pinnedFeatureValues` for the process, so availability is fixed at session start; installed mods load under `tengu_plugin_hooks_modules` (default true) | The 2.1.287 bundle on this machine (`var vr=()=>ft()&&lo(Ce(),at)`; `function lo(e,n){let r=Co().pinnedFeatureValues??=new Map;if(!r.has(e))r.set(e,C(e,n));...}`; `var HXe=\"tengu_plugin_hooks_modules\";var kjt=!0;var dQ=()=>C(HXe,kjt)`) and `jq .cachedGrowthBookFeatures ~\u002F.claude.json` | A release renames `cachedGrowthBookFeatures` or `pinnedFeatureValues`, drops the rollout flag string (`builtin_plugin_notes.mods_flag_in_bundle` goes false), or docs publish per-plugin availability | 2026-10-01, Claude Code 2.1.287 |\n| The tools reference table lists tools by exact name and is partial: 45 of 80 statically named tools are in it, and it keeps `TaskOutput`, which the binary only names in a retired-names list | `docs\u002Fen\u002Ftools-reference.md` parsed by `--docs` (46 rows) against a `--binary-only` run on this machine | The tools table restructures (the nested `tools` block goes `broken`) or the `undocumented` count reaches zero | 2026-09-29, Claude Code 2.1.285 |\n\nThe changelog at `https:\u002F\u002Fraw.githubusercontent.com\u002Fanthropics\u002Fclaude-code\u002Fmain\u002FCHANGELOG.md` is the\nfastest way to explain a diff between two runs; `--docs` attaches its per-name history. Commands\nappear and disappear between releases. Cite the changelog line that introduced or removed a\ncommand rather than asserting it changed.\n\n## Staying current as Claude Code ships\n\nThis skill reads Claude Code internals that change every release, so **drift must never be\nsilent**.\n\n**Read the integrity block before quoting any number, lane by lane.** Every run carries one, and it\nstates per lane (`builtin_commands`, `bundled_skills`, `plugin_backed`, `bundled_workflows`,\n`builtin_agents`, `builtin_tools`, `builtin_plugins`) whether that lane's counts are verified or merely believed; the top-level status is the worst lane.\n`docs_crosscheck` carries its own `status` (`ok`, `degraded`, `broken`, `unavailable`) and never\nchanges the integrity status or the self-check exit code:\n\n| Status | Means | Do |\n|---|---|---|\n| `ok` | Build matches the last validated release, every check passed | Report that lane's counts as totals |\n| `degraded` | Extraction worked, but something is unaccounted for | Report that lane's counts as **floors**, and say what is unaccounted for |\n| `broken` | A canary is missing or nothing resolved in that lane | Do not report that lane's counts at all, not even caveated or as the number of entries the file lists; name the lane, call it `broken`, and quote its `problems` text verbatim; the other lanes' counts stand |\n\nA single broken lane makes the top level `degraded`, not `broken`, so the other lanes' counts are\nnever withheld. Top-level `broken` means every lane is broken or the binary could not be read.\n\nThe dangerous failure is a clean, confident, short list from a new registration path, not a crash,\nso the checks catch shortfall: canary commands, a canary workflow, canary subagents and tools, a\nminimum ratio of resolved commands to registration tokens present, a sweep for registrar-shaped\nexports the script does not know about, the resolved-versus-seen gap on bundled skills, and\nunresolved subagent and tool names.\n\n**Run the drift check on a schedule, not on incident:**\n\n```bash\npython3 \"${CLAUDE_PLUGIN_ROOT}\u002Fskills\u002Finventory\u002Fscripts\u002Finventory.py\" --deps-dir \"${CLAUDE_PLUGIN_DATA}\" --self-check\n```\n\nIt prints one verdict line and exits `0` ok, `1` broken, `3` degraded; `2` is argparse's usage\nerror, never a degraded run. `\u002Fharness-ops:changelog apply` runs it after integrating a CLI\nrelease, with a `--docs` extraction it diffs against the previous run's, and files a work item when\nthe verdict is not `ok`.\n\n**What a maintainer updates.** Most releases need no change: registrar names are discovered and the\nbundle is found by export name. When a run goes `degraded` or `broken`, each verdict maps to one\nedit in `scripts\u002Finventory.py`, per the verdict table in\n[reference\u002Fextraction.md](reference\u002Fextraction.md). After revalidating against the new build, bump\n`VALIDATED_AGAINST` to that version; it is the one constant that turns \"believed\" back into\n\"verified\".\n\n**For downstream consumers.** An older plugin version against a newer CLI reports `degraded` or\n`broken` rather than a wrong answer; fixes ship as a new plugin version, which\n`\u002Fharness-ops:plugins sync` carries. Never quietly widen a count to make a status look better.\n\n## Next\n\n\u002Fharness-ops:audit-native-overlap\n\nIt reads this inventory's invocability, argument-hint, and workflow fields to find where local\nskills duplicate a native surface.\n\n## Gotchas\n\n- **A name in the build is not always a command.** Strings such as `alias` and `todos` match a\n  naive `name:\"…\"` search and are not commands. The brace-depth reader plus the `type:` requirement\n  is what excludes them. Verified 2026-09-29 against Claude Code 2.1.284, by running\n  `inventory.py --binary-only` on this machine: neither name appears under `builtin_commands`.\n  Recheck when the extractor's `type:` requirement changes or a release adds a command by either\n  name.\n- **An alias is not a separate command.** Count commands once and list aliases beside them, or\n  your total will drift from `\u002Fhelp`. Pointer: when stating which commands are aliases, fetch\n  [commands: all commands](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fcommands#all-commands) live; our probe,\n  a run of `inventory.py --binary-only` on Claude Code 2.1.284 (2026-09-29), reported `usage` with\n  aliases `cost` and `stats`. As of: 2026-11-10. Recheck trigger: the commands page changes the\n  alias rows for `\u002Fusage` or the extraction reports a different alias set for `usage`.\n- **A command in the docs can be a skill in the binary.** The commands page lists `\u002Fschedule` as a\n  command; the binary registers it as a bundled skill, so `docs_crosscheck` flags `kind_mismatch`.\n  Report the binary's kind, since it decides namespacing and model invocability. Verified\n  2026-09-29 against Claude Code 2.1.284 by a `--binary-only --docs` run on this machine. Recheck\n  when the page adds a **Skill** marker to `\u002Fschedule` or the binary registers it as a command.\n- **A bundled skill can also appear as a command object.** When a name registers as both, the skill\n  registration wins; the script drops the duplicate so one capability is not counted twice.\n- **An npm install has no embedded bundle.** The launcher script is small and loads its bundle\n  elsewhere. The script reports this rather than parsing a shim, and the disk half still works.\n- **The bundled-skills directory is a lazy cache.** Skills are extracted there on first use, so it\n  under-reports. It is never the source for the bundled-skill list.\n",{"metadata":24,"description":7,"argument-hint":28,"user-invocable":29,"disable-model-invocation":20},{"cadence":25,"summary":26,"workflow-stage":27},"weekly","Enumerate every command, skill, agent, and plugin component this machine can invoke","operator","[--builtin|--plugins|--bundled|--agents|--tools|--hooks|--docs] [--marketplace \u003Cm>] [--diff \u003Cf>]",true,"main","d0e68c3f36b92a1d6381f0ba0060d7a0b8b886c5","plugins\u002Fharness-ops\u002Fskills\u002Finventory",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":40,"prompt":50,"zipAvailable":29},[41,44,47],{"tool":42,"command":43},"npx","npx skills add melodic-software\u002Fclaude-code-plugins --skill inventory",{"tool":45,"command":46},"bunx","bunx skills add melodic-software\u002Fclaude-code-plugins --skill inventory",{"tool":48,"command":49},"pnpm","pnpm dlx skills add melodic-software\u002Fclaude-code-plugins --skill inventory","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002FSKILL.md",[52,58,62,68,72,76,80,84,88,92,96,100,104,108,112,116],{"path":53,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":56,"sourceHtmlUrl":57},"evals\u002Fevals.json","REFERENCE","参考文档",10826,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fevals\u002Fevals.json",{"path":59,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":60,"sourceHtmlUrl":61},"reference\u002Fextraction.md",66344,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Freference\u002Fextraction.md",{"path":63,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":66,"sourceHtmlUrl":67},"scripts\u002Fcompare_reports.py","SCRIPT","脚本",9225,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Fcompare_reports.py",{"path":69,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":70,"sourceHtmlUrl":71},"scripts\u002Fdocs_crosscheck.py",23553,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Fdocs_crosscheck.py",{"path":73,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":74,"sourceHtmlUrl":75},"scripts\u002Finventory.py",268926,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Finventory.py",{"path":77,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":78,"sourceHtmlUrl":79},"scripts\u002Fjs\u002Fpackage-lock.json",2666,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Fjs\u002Fpackage-lock.json",{"path":81,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":82,"sourceHtmlUrl":83},"scripts\u002Fjs\u002Fpackage.json",318,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Fjs\u002Fpackage.json",{"path":85,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":86,"sourceHtmlUrl":87},"scripts\u002Fjs\u002Fparser_helper.cjs",140317,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Fjs\u002Fparser_helper.cjs",{"path":89,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":90,"sourceHtmlUrl":91},"scripts\u002Fparser_reader.py",28937,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Fparser_reader.py",{"path":93,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":94,"sourceHtmlUrl":95},"scripts\u002Ftest_compare_reports.py",11911,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Ftest_compare_reports.py",{"path":97,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":98,"sourceHtmlUrl":99},"scripts\u002Ftest_fixture_parse.py",5769,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Ftest_fixture_parse.py",{"path":101,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":102,"sourceHtmlUrl":103},"scripts\u002Ftest_inventory_under_parser.py",2512,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Ftest_inventory_under_parser.py",{"path":105,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":106,"sourceHtmlUrl":107},"scripts\u002Ftest_inventory.py",178259,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Ftest_inventory.py",{"path":109,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":110,"sourceHtmlUrl":111},"scripts\u002Ftest_parser_reader.py",63901,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Ftest_parser_reader.py",{"path":113,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":114,"sourceHtmlUrl":115},"scripts\u002Ftest_reader_findings.py",82206,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Ftest_reader_findings.py",{"path":117,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":118,"sourceHtmlUrl":119},"scripts\u002Ftest_skills_index.py",9959,"https:\u002F\u002Fgithub.com\u002Fmelodic-software\u002Fclaude-code-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fharness-ops\u002Fskills\u002Finventory\u002Fscripts\u002Ftest_skills_index.py",[121,128,135,142,147,154],{"skillNo":122,"name":123,"nameZh":6,"description":124,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":125,"categories":126,"keywords":127,"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":129,"name":130,"nameZh":6,"description":131,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":132,"categories":133,"keywords":134,"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":136,"name":137,"nameZh":6,"description":138,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":139,"categories":140,"keywords":141,"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":143,"name":137,"nameZh":6,"description":138,"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":146,"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":148,"name":149,"nameZh":6,"description":150,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":151,"categories":152,"keywords":153,"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":155,"name":156,"nameZh":6,"description":157,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":158,"categories":159,"keywords":160,"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",[],[]]