feat(plugin): overhaul skills, hooks, and docs for Claude plugin v2.3.5 (#41)

* release: bump to v2.0.3

* fix: bump plugin version to 2.0.14 to bust marketplace cache

Aligns plugin version with MCP server v2.x major version line and
forces Claude Code to re-fetch updated marketplace listing (removes
stale clockwork/current dual-plugin view from cached installs).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* docs: overhaul CLAUDE.md for v3 accuracy and token efficiency

- Remove stale v2/Clockwork content (wrong tool names, wrong paths, wrong architecture)
- Update all source paths to current/ module (io.github.jpicklyk.mcptask.current)
- Remove MCP tool documentation (covered by tool descriptions at session start)
- Remove sections Claude can infer from code (design patterns, version format, etc.)
- Reduce from 443 lines (~5,500 tokens) to 161 lines (~1,700 tokens) — 69% reduction

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* docs: add per-project data isolation section to README

Explains how to scope the Docker volume name per project for isolated
SQLite databases, and recommends combining it with the per-project .mcp.json.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* feat(plugin): add dependency-manager, batch-complete, and note-viewer skills

Three new plugin skills covering the remaining MCP tool surface gaps:
- dependency-manager: visualize, create, delete, and diagnose dependencies
- batch-complete: bulk complete/cancel item trees with gate checking
- note-viewer: read, edit, and create notes with gate-aware reporting

Also polishes existing skills (create-item, schema-builder, status-progression,
work-summary) with expanded troubleshooting, argument-hints, and trigger phrases.
Renames skill files from skill.md to SKILL.md for consistency. Adds quick-start
skill and schema-builder reference examples. Bumps plugin version to 2.2.0.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* docs: rewrite quick-start for Claude Code users

Reframe from API reference to product walkthrough: conversational
examples first, tool calls second. Promote plugin from optional Step 5
to recommended Step 4. Add plan-mode pipeline diagram (Step 5) and
session resume section (Step 7). Fix note schema YAML to correct
map-key format. Update skills table to current 8 skills.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat(plugin): overhaul plugin skills and hooks for v2.3.5

- Extract pre-plan and post-plan hooks into internal skills
  (user-invocable: false) for maintainability
- Add definition floor concept to pre-plan workflow — agent reads
  config.yaml for schema discovery before writing plans
- Remove manage-notes skill (agent-only concern, not user-facing)
- Fill schema discovery gap in create-item (schema-aware tagging
  with default fallback) and post-plan-workflow (tag at materialization)
- Rename schema-builder to manage-schemas with starter templates
  (feature-implementation, bug-fix) for new user onboarding
- Remove note-viewer skill (superseded by status-progression)
- Add "schemas vs notes" callouts in quick-start and workflow-guide
  to clarify the distinction between config rules and DB content
- Update all cross-references and bump plugin version to 2.3.5

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: jpicklyk

.claude-plugin/marketplace.json

@@ -10,7 +10,7 @@
   "plugins": [
     {
       "name": "task-orchestrator",
-      "version": "2.0.14",
+      "version": "2.3.5",
       "description": "Skills, hooks, and workflows for MCP Task Orchestrator (v3). Schema-aware context, note-driven workflow, and thin session hooks.",
       "author": {
         "name": "Jeff Picklyk",

README.md

@@ -144,8 +144,11 @@ After installing, restart Claude Code and verify with `/plugin list` — you sho
 |---------|-------------|
 | `/task-orchestrator:work-summary` | Insight-driven dashboard: active work, blockers, and next actions |
 | `/task-orchestrator:create-item` | Create a tracked work item from the current conversation context |
+| `/task-orchestrator:quick-start` | Interactive onboarding — teaches by doing, adapts to empty or populated workspaces |
+| `/task-orchestrator:manage-schemas` | Create, view, edit, delete, and validate note schemas in config |
 | `/task-orchestrator:status-progression` | Navigate role transitions; shows gate status and the correct trigger |
-| `/task-orchestrator:schema-builder` | Interactively design a note schema for a new work item type |
+| `/task-orchestrator:dependency-manager` | Visualize, create, and diagnose dependencies between work items |
+| `/task-orchestrator:batch-complete` | Complete or cancel multiple items at once — close out features or workstreams |
 
 **Hooks** — automatic, no invocation needed:
 

claude-plugins/CLAUDE.md

@@ -37,7 +37,7 @@ Use semantic versioning (`major.minor.patch`):
 
 | Plugin | Directory | Current Version |
 |--------|-----------|-----------------|
-| `task-orchestrator` | `claude-plugins/task-orchestrator/` | `2.0.14` |
+| `task-orchestrator` | `claude-plugins/task-orchestrator/` | `2.3.5` |
 
 > Update this table when versions change.
 

claude-plugins/task-orchestrator/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "task-orchestrator",
-  "version": "2.0.14",
+  "version": "2.3.5",
   "description": "Claude Code integration for Current (v3) MCP Task Orchestrator — thin hooks, schema-aware context, note-driven workflow",
   "skills": "./skills",
   "hooks": "./hooks/hooks-config.json",

claude-plugins/task-orchestrator/hooks/post-plan.mjs

@@ -1,17 +1,11 @@
 #!/usr/bin/env node
-// PostToolUse ExitPlanMode — injects post-approval reminder
+// PostToolUse ExitPlanMode — triggers the post-plan-workflow skill
 const output = {
   hookSpecificOutput: {
     hookEventName: "PostToolUse",
-    additionalContext: `## Plan Approved — Current (v3) Next Steps
+    additionalContext: `## Plan Approved — MCP Task Orchestrator
 
-Plan approval is the green light for the full pipeline. Proceed without stopping:
-
-1. **Materialize:** Create MCP containers (items, dependencies) — complete before implementation
-2. **Implement:** Dispatch subagents with MCP item UUIDs. Each agent must call advance_item(trigger="start") at the start and advance_item(trigger="complete") at the end.
-3. **Verify:** Run \`get_context()\` health check after all agents complete.
-
-Do NOT use AskUserQuestion between phases — proceed autonomously.`
+Invoke the \`task-orchestrator:post-plan-workflow\` skill for the full post-approval workflow. It covers materialization, implementation dispatch, and verification.`
   }
 };
 process.stdout.write(JSON.stringify(output));

claude-plugins/task-orchestrator/hooks/pre-plan.mjs

@@ -1,22 +1,11 @@
 #!/usr/bin/env node
-// PreToolUse EnterPlanMode — injects planning workflow guidance
+// PreToolUse EnterPlanMode — triggers the pre-plan-workflow skill
 const output = {
   hookSpecificOutput: {
     hookEventName: "PreToolUse",
-    additionalContext: `## Planning Mode — Current (v3) Workflow
+    additionalContext: `## Planning Mode — MCP Task Orchestrator
 
-Before implementing, call \`get_context(itemId=<activeItemId>)\` to:
-- See the note schema for this item's tags
-- Check gate status (which required notes must be filled)
-- Get the guidance pointer for authoring notes
-
-**Planning workflow:**
-1. Create items with \`manage_items\` or \`create_work_tree\`
-2. Check \`get_context\` for expected notes per item
-3. Fill required notes with \`manage_notes\`
-4. Advance with \`advance_item(trigger="start")\` — gate enforcement will verify notes
-
-**After plan approval:** Materialize containers, then dispatch implementation agents with MCP item UUIDs.`
+Invoke the \`task-orchestrator:pre-plan-workflow\` skill for the full planning workflow. It will guide you through checking existing MCP state to set the definition floor before writing your plan.`
   }
 };
 process.stdout.write(JSON.stringify(output));