fix(claude-ops): correct inventory field bleed, manifest paths, and scope coverage

plugins/claude-ops/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
   "name": "claude-ops",
-  "version": "0.31.0",
+  "version": "0.31.1",
   "description": "Claude Code operations toolkit. Ten skills: inventory (read-only enumeration of the complete invocable surface \u2014 every built-in CLI command with aliases and hidden/gated status, every bundled skill, and every component of every installed plugin across all marketplaces; reads the shipped binary because upstream publishes no built-in command list, and carries an integrity verdict so a drifted build reports counts as floors rather than silently short totals), audit-install-state (read-only audit of the machine-scope ~/.claude installation directory and ~/.claude.json \u2014 full inventory split into an authored surface and rolled-up bulk trees, product-managed retention vs genuinely unmanaged state, filename-scheme resolution before any process-liveness check, and deliberate/mid-experiment detection; reports, never deletes), audit-performance (read-only slowness-diagnostic capture run at the moment the machine or a session feels slow \u2014 CLI version, retention-sweep health including the silent unparsable-settings pause, a timed census walk of the install tree as a sweep-cost proxy, active-session and plugin-fleet counts, a process census, and a bundled known-performance-issues reference; separates the three documented suspects \u2014 accumulated state, version regression, component bloat \u2014 and routes remediation out; reports, never mutates), observability (read locally captured telemetry \u2014 OTEL store, collector, hook-event JSONL, ccusage \u2014 with trend reports and store pruning), known-issues (search known Claude product GitHub bugs, check service health, maintain a persistent tracked-issue registry), changelog (ingest Claude Code changelog entries and integrate them into the current repo), plugins (bring a machine's plugin fleet current on demand \u2014 marketplace refresh, effective-scope updates including in-repo project/local installs, new-plugin install per policy, scope-divergence detection and explicit convergence), morning-brief (read-only gh-based operator morning view \u2014 queue-label counts, merge-ready PRs, parked decisions with their RECOMMENDED lines, and loop-lane telemetry freshness), lanes (start/restart/stop/status loop lanes as named background Claude Code sessions seeded from canonical prompt files, with per-lane model/effort, a repo-pull + marketplace-refresh launch step, and a consume-restarts action \u2014 an OS-schedulable reader that relaunches stopped lanes whose telemetry carries a restart_request), and a re-runnable setup action that settles where the known-issues registry lives. Plus a family of seven advisory *-audit telemetry-emitter hooks (API errors, config changes, instruction loads, permission denials, pre-compaction, skill usage, tool failures) that emit the shared hook-telemetry envelope, and a reference sink that maps envelopes into the hook-events.jsonl the observability skill reads.",
   "author": {
     "name": "Melodic Software",

plugins/claude-ops/CHANGELOG.md

@@ -3,6 +3,38 @@
 All notable changes to the `claude-ops` plugin are documented here. Format follows
 [Keep a Changelog](https://keepachangelog.com/en/1.1.0/); this plugin uses semantic versioning.
 
+## [0.31.1]
+
+### Fixed
+
+- **`skills/inventory` bundled-skill fields could bleed from the next registration.** The extractor
+  read each registration through a fixed 4000-character window — the failure mode `build_brace_map`
+  exists to prevent for commands, and the one `reference/extraction.md` names as the thing not to
+  do. A registration omitting a description adopted the following one's. Fields are now bound to
+  their own literal via the brace map, and an unmatched brace is counted and surfaced rather than
+  silently skipped.
+- **Manifest-declared component paths were ignored.** `PLUGIN_COMPONENTS` carried a manifest key per
+  component and a comment claiming the manifest is read before the tree; nothing read it. A declared
+  path replaces the default directory, so scanning defaults regardless reported components a plugin
+  does not ship. Dotted keys resolve the `experimental` block.
+- **`--self-check` lost its diagnostic when no binary was found.** `pick_binary` stores its
+  explanation under `reason`; the self-check path read only `error` and printed a generic message.
+- **An unreadable CLI version passed silently.** It is itself a drift signal, so it now degrades the
+  verdict instead of skipping the comparison.
+- **`--self-check` degraded and an argparse usage error both exited 2.** A CI gate treating 2 as
+  "degraded, warn" would silently swallow a mistyped flag. Degraded is now 3, leaving 2 to argparse:
+  0 ok, 1 broken, 2 usage error, 3 degraded.
+
+### Added
+
+- **`skills/inventory` reads installed plugins and project scope.** Only marketplace catalogs were
+  scanned, so a plugin installed from a marketplace that is no longer cached was invisible;
+  `disk.installed_plugins` now walks the plugin cache, and catalog, installed, and enabled are
+  reported as three distinct sets. A project's `.claude` tree contributes skills, agents, and wired
+  hook events that no machine-scope scan sees — `--project-dir` defaults to the working directory.
+  Wired hook events are reported, never hook scripts on disk, which would repeat the
+  present-versus-active error the skill warns about.
+
 ## [0.31.0]
 
 ### Added

plugins/claude-ops/skills/inventory/SKILL.md

@@ -91,7 +91,11 @@ One per line, alphabetical, with aliases and hidden/gated markers.
 One per line, alphabetical.
 
 ## Plugin components
-Grouped by marketplace, then plugin, with a component-type breakdown.
+Installed plugins grouped by marketplace, then plugin, with a component-type breakdown.
+Catalog-only entries — offered by a cached marketplace but not installed — are listed separately.
+
+## Project scope
+Skills, agents, and wired hook events from the current project's `.claude` tree, when present.
 
 ## Provenance
 Which source produced which section, and anything the run could not resolve.
@@ -119,8 +123,14 @@ hooks live in its own manifest. Report the map as read and route the verdict to
 alongside `resolved`. When they differ, some registration used a dynamically computed name; say so
 rather than reporting the smaller number as complete.
 
-**A marketplace checkout is not an installation.** Plugins under a cached marketplace are a catalog
-of what is *available*. Only `enabledPlugins` says what loads.
+**A marketplace checkout is not an installation, and neither is enablement.** Three different sets:
+a cached marketplace is a catalog of what is *available*, `disk.installed_plugins` is what is
+*present locally*, and `enabledPlugins` governs what *loads*. They routinely disagree — report the
+one the question is actually about, and say which you used.
+
+**A hook script on disk is not a wired hook.** Project scope reports hook *events* declared in
+settings, not files sitting in a `.claude/hooks/` directory. A script nothing references is dead
+weight, and listing it as a hook repeats the same present-versus-active error.
 
 ## How the binary read works
 
@@ -177,8 +187,9 @@ exports the script does not know about, and the resolved-versus-seen gap on bund
 python3 "${CLAUDE_PLUGIN_ROOT}/skills/inventory/scripts/inventory.py" --self-check
 ```
 
-It prints one verdict line and exits `0` ok, `1` broken, `2` degraded — so it works as a CI gate, a
-loop-lane step, or a post-update check without parsing JSON. The natural trigger is a CLI release:
+It prints one verdict line and exits `0` ok, `1` broken, `3` degraded — so it works as a CI gate, a
+loop-lane step, or a post-update check without parsing JSON. `2` is left to argparse for a usage
+error, so a mistyped flag can never be mistaken for a degraded run. The natural trigger is a CLI release:
 `/claude-ops:changelog` already ingests those, and this is the check to run when it reports one.
 
 **What a maintainer actually updates.** Most releases need no change — registrar names are