feat(claude-ops): add plugins skill for plugin-fleet sync/audit/converge

README.md

@@ -80,7 +80,7 @@ Browse and manage with `/plugin`. To refresh after updates: `/plugin marketplace
 - [`playbooks`](plugins/playbooks) — Doctrine and knowledge playbooks as on-demand skills, plus a maintainer-facing update skill. boris — Boris Cherny's Claude Code workflow tips (howborisusesclaudecode.com); thariq — Anthropic's internal skill-authoring playbook; fable-5 — Claude Fable 5's operating doctrine (self-authored, no upstream). The boris and thariq packs vendor a verbatim upstream baseline; /playbooks:update drift-checks and syncs those baselines centrally (maintainers).
 - [`claude-config`](plugins/claude-config) — Three audit skills for a repo's Claude Code configuration: audit (settings.json / .mcp.json / hooks / plugins / permissions drift), automation-gaps (evidence-gated verdicts on automation gaps), and permission-hygiene (allow-rule / allowed-tools grants for auto-mode durability and portability).
 - [`claude-memory`](plugins/claude-memory) — Audits the Claude Code instruction/memory layer — CLAUDE.md, CLAUDE.local.md, .claude/rules/, and auto-memory — against a checklist derived from official Claude Code documentation. A deterministic script-backed spine (MEMORY.md index integrity, orphan always-loaded rules) yields identical findings on identical repo state; judgment-tier checks apply fixed criteria with model reading. Actions: audit (default), fix (per-item approval), update (refresh criteria from current docs), report.
-- [`claude-ops`](plugins/claude-ops) — Claude Code operations toolkit. Four skills: observability (read locally captured telemetry — OTEL store, collector, hook-event JSONL, ccusage — with trend reports and store pruning), troubleshoot (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), and a re-runnable setup action that settles where the troubleshooting 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.
+- [`claude-ops`](plugins/claude-ops) — Claude Code operations toolkit. Five skills: observability (read locally captured telemetry — OTEL store, collector, hook-event JSONL, ccusage — with trend reports and store pruning), troubleshoot (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 — marketplace refresh, effective-scope updates including in-repo project/local installs, new-plugin install per policy, scope-divergence detection and explicit convergence), and a re-runnable setup action that settles where the troubleshooting 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.
 - [`skill-quality`](plugins/skill-quality) — Skill-authoring QA tooling: a static contract checker that runs seventeen deterministic checks over a Claude Code skill (frontmatter, listing-budget cap, trigger-keyword preservation, line caps, broken internal refs, markdownlint, gotchas surface, evals presence) and a bundled evals.json schema for validation. Runs against any repo's skills directory via the convention-resolution ladder — no baked layout.
 
 ### Security

docs/topics/plugin-fleet-sync-skill/design/design-resolution.md

@@ -0,0 +1,39 @@
+# Design resolution — plugin-fleet-sync-skill
+
+outcome: early-exit (Tier B)
+
+Reason: the contract surface was fully resolved by the /interview Brief (PLAN.md `## Brief`,
+Decisions 1–8) backed by three verified research passes (official CC docs, repo conventions,
+scope-precedence empirics). No new code types or module topology — the deliverable is a
+markdown skill orchestrating the `claude plugin` CLI plus one read-only state-inspection
+script. Type surface is small enough to sketch here; full /design would re-derive the Brief.
+
+## Type sketch
+
+**Action surface** (router in SKILL.md):
+
+- `sync [marketplace|all]` — default; mutating via CLI only
+- `audit [marketplace|all]` — read-only dry-run of sync + converge predictions
+- `converge` — explicit scope consolidation; only committed-settings-touching action
+
+**userConfig** (plugin.json, personal policy scalar):
+
+- `install_new`: `ask` (default) | `all` | `none` — enum support in the manifest schema is a
+  Phase-1 verification; fallback is a string field validated in skill prose.
+
+**State-inspection output** (read-only script → JSON consumed by the skill):
+
+```json
+{
+  "marketplace": {"name": "...", "autoUpdate": false, "lastUpdated": "..."},
+  "catalog": ["plugin", "..."],
+  "installed": [{"id": "plugin@mp", "scope": "user|project|local", "version": "...", "projectPath": "..."}],
+  "enabled": {"plugin@mp": true},
+  "missing_from_install": ["..."],
+  "missing_from_enabled": ["..."],
+  "divergences": [{"id": "...", "scopes": [{"scope": "...", "version": "...", "projectPath": "..."}]}]
+}
+```
+
+Inputs are internal CC state files read-only (`installed_plugins.json`, `known_marketplaces.json`,
+marketplace `marketplace.json`, settings `enabledPlugins`); every mutation goes through the CLI.

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

@@ -1,14 +1,14 @@
 {
   "$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
   "name": "claude-ops",
-  "version": "0.8.0",
-  "description": "Claude Code operations toolkit. Four skills: observability (read locally captured telemetry — OTEL store, collector, hook-event JSONL, ccusage — with trend reports and store pruning), troubleshoot (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), and a re-runnable setup action that settles where the troubleshooting 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.",
+  "version": "0.9.0",
+  "description": "Claude Code operations toolkit. Five skills: observability (read locally captured telemetry — OTEL store, collector, hook-event JSONL, ccusage — with trend reports and store pruning), troubleshoot (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 — marketplace refresh, effective-scope updates including in-repo project/local installs, new-plugin install per policy, scope-divergence detection and explicit convergence), and a re-runnable setup action that settles where the troubleshooting 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",
     "email": "info@melodicsoftware.com"
   },
   "license": "MIT",
-  "keywords": ["claude-code", "operations", "observability", "otel", "troubleshooting", "changelog", "monitoring", "hooks", "telemetry", "audit"],
+  "keywords": ["claude-code", "operations", "observability", "otel", "troubleshooting", "changelog", "monitoring", "hooks", "telemetry", "audit", "plugins", "marketplace"],
   "userConfig": {
     "registry_dir": {
       "type": "string",
@@ -19,6 +19,11 @@
       "type": "string",
       "title": "Skill-usage log directory (project-relative)",
       "description": "Optional contained project-relative directory where the skill-usage-audit hook writes skill-usage.jsonl. Absolute, drive, UNC, traversal, and escaping-symlink paths are invalid. Leave unset to use .claude/observability."
+    },
+    "install_new": {
+      "type": "string",
+      "title": "New-plugin install policy for the plugins skill's sync action",
+      "description": "Controls what `sync` does with catalog plugins that aren't installed yet. Valid values: \"ask\" (default — offer them in one batched multi-select prompt), \"all\" (install every one automatically), \"none\" (report only, never install). The manifest schema has no enum type, so this validates in prose, not JSON Schema; any other value is treated as \"ask\"."
     }
   }
 }

plugins/claude-ops/CHANGELOG.md

@@ -3,6 +3,21 @@
 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.
 
+## [0.9.0]
+
+### Added
+
+- New `plugins` skill (`/claude-ops:plugins`): brings a machine's plugin fleet current on demand.
+  `sync` (default) refreshes marketplaces, updates in-repo project/local-scope installs plus the
+  user-scope sweep, installs new catalog plugins per the `install_new` policy, and fills any
+  `enabledPlugins` completeness gap — all CLI-mediated, never hand-editing Claude Code's internal
+  state files. `audit` runs the same algorithm read-only. `converge` is the one action that can
+  touch a committed `.claude/settings.json`: it detects actionable (version-behind) scope
+  divergence, previews and confirms per plugin, then surfaces the resulting diff for review — never
+  auto-committed, and it aborts outright in an autonomous session. Adds a read-only
+  `scripts/fleet-state.sh` state-inspection script and the `install_new` userConfig scalar
+  (`ask` default / `all` / `none`).
+
 ## [0.8.0]
 
 ### Changed

plugins/claude-ops/README.md

@@ -1,12 +1,13 @@
 # claude-ops
 
 A Claude Code plugin for running Claude Code well over time — one cohesive
-capability across four skills and a family of telemetry-emitter hooks.
+capability across five skills and a family of telemetry-emitter hooks.
 Observability reads what your sessions actually did, troubleshooting tracks what
 upstream has broken, changelog integration keeps your repo current with what
-upstream has shipped, a re-runnable `setup` action settles where the
-troubleshooting registry lives, and the `*-audit` hooks feed observability with
-per-hook execution telemetry Claude Code's native OTEL cannot see.
+upstream has shipped, the plugins skill keeps your own plugin fleet current, a
+re-runnable `setup` action settles where the troubleshooting registry lives,
+and the `*-audit` hooks feed observability with per-hook execution telemetry
+Claude Code's native OTEL cannot see.
 
 ## Skills
 
@@ -15,6 +16,7 @@ per-hook execution telemetry Claude Code's native OTEL cannot see.
 | `/claude-ops:observability` | Reads locally captured Claude Code telemetry — OTEL DuckDB store, machine-owned collector, optional Aspire dashboard, hook-event JSONL, ccusage — and renders cross-session trend reports (`session`/`day`/`week`/`month`/`since:`/`all` scopes). Read-only except the explicit `clean` action, which prunes the JSONL log and OTEL store by age. |
 | `/claude-ops:troubleshoot` | Searches known Claude product GitHub bugs before you build on a feature, checks service health and model quality, and maintains a persistent registry of tracked issues (what they block, workarounds, follow-ups when fixed). Actions: `status` (default), `search`, `check-all`, `scan`, `list`, `quality`, `create`. |
 | `/claude-ops:changelog` | Ingests Claude Code changelog entries and integrates them into the current repo: `fetch` (read-only display), `diff` (impact triage, no edits), `status` (applied versions from git history), and `apply` (full explore → research → interview → implement pipeline, explicit user intent only). |
+| `/claude-ops:plugins` | Brings a machine's plugin fleet current on demand: marketplace refresh, updates for the plugins that actually load (including in-repo project/local-scope installs), new-catalog-plugin install per policy, and scope-divergence detection. Actions: `sync` (default, CLI-mediated mutations only), `audit` (read-only dry run), `converge` (the one action that can touch a committed `.claude/settings.json` — previews and confirms per plugin first). |
 | `/claude-ops:setup` | Validates the troubleshooting-registry and skill-usage-log destinations, including path containment, and routes personal option changes through Claude Code's plugin configuration prompt. |
 
 ## The audit hooks
@@ -120,8 +122,13 @@ instead of failing.
 
 ## Configuration
 
-Two `userConfig` options:
+Three `userConfig` options:
 
+- **`install_new`** (string, optional) — new-catalog-plugin install policy for the `plugins`
+  skill's `sync` action. `ask` (default) offers not-yet-installed catalog plugins in one batched
+  multi-select prompt; `all` installs every one automatically; `none` reports them without
+  installing. The manifest schema has no `enum` type, so this validates in prose, not JSON Schema;
+  any other value is treated as `ask`.
 - **`registry_dir`** (string, optional) — project-relative directory for the
   troubleshoot issue registry (`registry.json`). Set it to keep the
   registry inside your repo (git-tracked, team-shared) instead of the