claude-autopm 3.23.0 → 3.23.1

package/autopm/.claude/commands/pm:epic-decompose.md

@@ -321,66 +321,84 @@ Creating task 002 (Database schema):
 ✓ Created
 ```
 
-### 4. Task File Format with Frontmatter
+### 4. MANDATORY: Read Issue Template Before Creating Tasks
+
+**CRITICAL**: Before creating ANY task, read the issue template:
+```
+Read .claude/templates/issue.xml
+Read .claude/rules/issue-structure.xml
+```
+
+Every task file MUST include ALL required sections from `issue.xml`. Missing sections = auto-reject.
+
+### 5. Task File Format with Frontmatter
 For each task, create a file with this exact structure:
 
 ```markdown
 ---
 name: [Task Title]
 status: open
+priority: [critical/high/medium/low]
+effort: [S/M/L/XL]
 created: [Current ISO date/time]
 updated: [Current ISO date/time]
-assigned_agent: .claude/agents/{category}/{agent-name}.md  # Specialized agent for this task
-agent_context:  # Optional: Agent-specific configuration
+assigned_agent: .claude/agents/{category}/{agent-name}.md
+agent_context:
   framework: [framework_name]
   language: [language_name]
-  approach: [implementation_approach]
 github: [Will be updated when synced to GitHub]
-depends_on: []  # List of task numbers this depends on, e.g., [001, 002]
-parallel: true  # Can this run in parallel with other tasks?
-conflicts_with: []  # Tasks that modify same files, e.g., [003, 004]
+depends_on: []
+parallel: true
+conflicts_with: []
 ---
 
 # Task: [Task Title]
 
-## Description
-Clear, concise description of what needs to be done
-
-## ⚠️ TDD Requirements
-**This project uses Test-Driven Development. You MUST:**
-1. 🔴 RED: Write failing test first
-2. 🟢 GREEN: Write minimal code to make test pass
-3. 🔵 REFACTOR: Clean up code while keeping tests green
+## Goal
+[One sentence: WHAT this task delivers and WHY it matters]
 
-See `.claude/rules/tdd.enforcement.md` for complete requirements.
+## Context
+[WHY this task exists — what triggered it, what currently works/doesn't,
+ what must be in place first. Include specific file paths, URLs, services.]
 
 ## Acceptance Criteria
-- [ ] Specific criterion 1
-- [ ] Specific criterion 2
-- [ ] Specific criterion 3
+- [ ] Specific, measurable criterion 1
+- [ ] Specific, measurable criterion 2
+- [ ] Specific, measurable criterion 3
 
-## Technical Details
-- Implementation approach
-- Key considerations
-- Code locations/files affected
+## Affected Files
+- `path/to/file.ts` — [what changes]
+- `path/to/test.ts` — [new test file]
 
-## Dependencies
-- [ ] Task/Issue dependencies
-- [ ] External dependencies
+## Related Issues
+- depends on #NNN — [why]
+- blocks #NNN — [why]
+- (or "None" if standalone)
 
-## Effort Estimate
-- Size: XS/S/M/L/XL
-- Hours: estimated hours
-- Parallel: true/false (can run in parallel with other tasks)
+## Verification Steps
+[Concrete commands to verify this task is done:]
+- `npm test -- --filter=feature` → all pass
+- `curl http://localhost:3000/api/endpoint` → 200
+- `kubectl get pods -n app` → Running
 
 ## Definition of Done
 - [ ] Tests written FIRST (RED phase)
 - [ ] Code implemented (GREEN phase)
 - [ ] Code refactored (REFACTOR phase)
 - [ ] All tests passing
-- [ ] Documentation updated
-- [ ] Code reviewed
-- [ ] Deployed to staging
+- [ ] Verification steps confirmed
+- [ ] PR approved and merged
+
+## ⚠️ TDD Requirements
+1. 🔴 RED: Write failing test first
+2. 🟢 GREEN: Write minimal code to make test pass
+3. 🔵 REFACTOR: Clean up code while keeping tests green
+
+## Technical Constraints
+[Stack, patterns, prohibitions — optional but recommended]
+
+## Out of Scope
+[What NOT to touch — prevents AI from "improving" neighbors]
 ```
 
 ### 3. Task Naming Convention

package/autopm/.claude/commands/pm:prd-new.md

@@ -356,18 +356,50 @@ Parse the arguments to detect the mode:
    - [Internal systems or components]
    - [Third-party libraries or tools]
 
-   ## Implementation Plan
-   ### Phase 1: Foundation
-   - [ ] [Initial tasks]
+   ## Architecture Design
+   ### Component Tree
+   [Map out the key new files/modules/components that will be created:]
+   ```
+   src/
+   ├── feature/
+   │   ├── Component.tsx
+   │   ├── hooks/useFeature.ts
+   │   └── utils/helpers.ts
+   ```
 
-   ### Phase 2: Core Features
-   - [ ] [Main development]
+   ### Data Flow
+   [How data moves through the system for this feature:]
+   - User action → API endpoint → Service → Database → Response
 
-   ### Phase 3: Enhancement
-   - [ ] [Additional features]
+   ### State Management
+   [New state needed — stores, context, caching strategy]
 
-   ### Phase 4: Release
-   - [ ] [Final testing and deployment]
+   ## Implementation Plan
+   ### Phase 1: Analyze (before writing code)
+   - [ ] Read existing codebase patterns
+   - [ ] Identify integration points
+   - [ ] Confirm architecture with stakeholder
+
+   ### Phase 2: Foundation
+   - [ ] Database migrations / models
+   - [ ] Core API endpoints (backend)
+   - [ ] Base component scaffolding (frontend)
+
+   ### Phase 3: Core Features
+   - [ ] [Main feature implementation]
+   - [ ] [Integration with existing systems]
+
+   ### Phase 4: Polish & Release
+   - [ ] [Testing — unit, integration, e2e]
+   - [ ] [Documentation]
+   - [ ] [Deployment]
+
+   ### Quick Win Path
+   [Minimum viable subset that works end-to-end in 1-3 days:]
+   - Day 1: [smallest vertical slice]
+   - Day 2: [connect to real data]
+   - Day 3: [basic UI showing results]
+   [This path gives a working demo before full implementation]
 
    ## Risks and Mitigation
    ### Technical Risks

package/autopm/.claude/rules/issue-structure.xml

@@ -6,13 +6,18 @@
 
   <template>
     <section name="goal" required="true">One sentence: WHAT and WHY</section>
-    <section name="context" required="true">Architecture location, files, dependencies</section>
+    <section name="context" required="true">
+      WHY this issue exists — problem background, current state, dependencies, environment.
+      Must answer: what triggered this, what works/doesn't, what must exist first.
+    </section>
     <section name="acceptance-criteria" required="true">Measurable checkboxes, not prose</section>
-    <section name="technical-constraints" required="false">Stack, patterns, prohibitions</section>
     <section name="affected-files" required="true">Exact file paths with change description</section>
+    <section name="related-issues" required="true">depends on / blocks / related to #NNN, or "None"</section>
+    <section name="verification" required="true">Concrete commands to verify (curl, kubectl, npm test, etc.)</section>
+    <section name="definition-of-done" required="true">Quality gate checkboxes — tests, deploy, PR merged</section>
+    <section name="constraints" required="false">Stack, patterns, prohibitions</section>
     <section name="out-of-scope" required="false">What NOT to touch</section>
     <section name="suggested-approach" required="false">Brief implementation strategy</section>
-    <section name="definition-of-done" required="false">Quality gates (tests, lint, coverage)</section>
   </template>
 
   <title-convention>
@@ -40,8 +45,12 @@
   </key-principles>
 
   <violations enforcement="auto-reject">
+    <violation>Creating issues without Context section (why it exists, current state)</violation>
     <violation>Creating issues without Acceptance Criteria</violation>
     <violation>Creating issues without Affected Files section</violation>
+    <violation>Creating issues without Verification Steps</violation>
+    <violation>Creating issues without Definition of Done</violation>
     <violation>Vague goal without concrete deliverable</violation>
+    <violation>Missing Related Issues (must list links or "None")</violation>
   </violations>
 </rule>

package/autopm/.claude/scripts/pm/dashboard-serve.js

@@ -400,7 +400,7 @@ function renderHTML() {
     border: 1px solid #30363d; border-radius: 8px; overflow: hidden;
   }
   .diagram-split .code-pane {
-    flex: 1; display: flex; flex-direction: column; border-right: 1px solid #30363d; min-width: 0;
+    width: 30%; display: flex; flex-direction: column; min-width: 150px; flex-shrink: 0;
   }
   .diagram-split .code-pane .pane-header {
     background: #161b22; padding: 6px 12px; font-size: 11px; color: #8b949e;
@@ -411,8 +411,13 @@ function renderHTML() {
     font-family: 'SF Mono', 'Fira Code', 'Cascadia Code', monospace; font-size: 13px;
     line-height: 1.5; resize: none; outline: none; tab-size: 2;
   }
+  .diagram-split .split-handle {
+    width: 5px; cursor: col-resize; background: #30363d; flex-shrink: 0;
+    transition: background 0.15s;
+  }
+  .diagram-split .split-handle:hover, .diagram-split .split-handle.dragging { background: #58a6ff; }
   .diagram-split .preview-pane {
-    flex: 1; display: flex; flex-direction: column; background: #161b22; min-width: 0;
+    flex: 1; display: flex; flex-direction: column; background: #161b22; min-width: 200px;
   }
   .diagram-split .preview-pane .pane-header {
     background: #161b22; padding: 6px 12px; font-size: 11px; color: #8b949e;
@@ -618,10 +623,11 @@ function renderHTML() {
         </div>
       </div>
       <div class="diagram-split" id="diagram-split">
-        <div class="code-pane">
+        <div class="code-pane" id="code-pane">
           <div class="pane-header"><span>Mermaid Source</span></div>
           <textarea id="diagram-code" spellcheck="false"></textarea>
         </div>
+        <div class="split-handle" id="split-handle"></div>
         <div class="preview-pane">
           <div class="pane-header">
             <span>Preview</span>
@@ -863,6 +869,35 @@ function resetPanZoom() {
   }
 }
 
+// Resizable split handle
+(function() {
+  const handle = document.getElementById('split-handle');
+  const codePane = document.getElementById('code-pane');
+  const split = document.getElementById('diagram-split');
+  if (!handle || !codePane || !split) return;
+  let dragging = false;
+  handle.addEventListener('mousedown', function(e) {
+    e.preventDefault();
+    dragging = true;
+    handle.classList.add('dragging');
+    document.body.style.cursor = 'col-resize';
+    document.body.style.userSelect = 'none';
+  });
+  document.addEventListener('mousemove', function(e) {
+    if (!dragging) return;
+    const rect = split.getBoundingClientRect();
+    const pct = Math.min(70, Math.max(10, ((e.clientX - rect.left) / rect.width) * 100));
+    codePane.style.width = pct + '%';
+  });
+  document.addEventListener('mouseup', function() {
+    if (!dragging) return;
+    dragging = false;
+    handle.classList.remove('dragging');
+    document.body.style.cursor = '';
+    document.body.style.userSelect = '';
+  });
+})();
+
 // Live preview on code change
 document.addEventListener('DOMContentLoaded', function() {
   const codeEl = document.getElementById('diagram-code');

package/autopm/.claude/templates/issue.xml

@@ -7,14 +7,42 @@
     <field name="assignee" type="string" default=""/>
     <field name="created" type="datetime" auto="true"/>
     <field name="updated" type="datetime" auto="true"/>
+    <field name="priority" type="enum" values="critical,high,medium,low" default="medium"/>
+    <field name="effort" type="enum" values="S,M,L,XL" default="M"/>
     <field name="started" type="datetime"/>
     <field name="completed" type="datetime"/>
   </frontmatter>
 
   <sections>
     <section name="goal" heading="Goal" required="true">One sentence: WHAT and WHY</section>
+    <section name="context" heading="Context" required="true">
+      Why this issue exists. Include:
+      - Problem background — what triggered this (bug report, user request, architectural gap)
+      - Current state — what works now, what doesn't, relevant infrastructure/config
+      - Dependencies — what must exist first (images, services, DNS, other issues)
+      - Access/environment — how to reproduce, where to verify (cluster, URL, CI)
+    </section>
     <section name="acceptance-criteria" heading="Acceptance Criteria" required="true">Measurable checkboxes</section>
     <section name="affected-files" heading="Affected Files" required="true">Exact file paths with change description</section>
+    <section name="related-issues" heading="Related Issues" required="true">
+      Links to blocking/blocked/related issues:
+      - depends on #NNN — must be done first
+      - blocks #NNN — this unblocks other work
+      - related to #NNN — shares context
+      - Write "None" if truly standalone
+    </section>
+    <section name="verification" heading="Verification Steps" required="true">
+      Concrete commands to verify the work is done. Examples:
+      - curl https://app.example.com → 200
+      - kubectl get pods -n app → Running
+      - npm test -- --filter=feature → pass
+    </section>
+    <section name="definition-of-done" heading="Definition of Done" required="true">
+      Checklist of quality gates that MUST pass before closing:
+      - [ ] Tests pass (unit, integration)
+      - [ ] Deployed to staging/verified
+      - [ ] PR approved and merged
+    </section>
     <section name="constraints" heading="Technical Constraints" required="false">Stack, patterns, prohibitions</section>
     <section name="out-of-scope" heading="Out of Scope" required="false">What NOT to touch</section>
     <section name="approach" heading="Suggested Approach" required="false">Brief implementation strategy</section>
@@ -22,6 +50,11 @@
 
   <validation>
     <rule>goal must be non-empty</rule>
+    <rule>context must be non-empty</rule>
+    <rule>acceptance-criteria must be non-empty</rule>
+    <rule>related-issues must be non-empty</rule>
+    <rule>verification must be non-empty</rule>
+    <rule>definition-of-done must be non-empty</rule>
     <rule>at least 1 acceptance criterion checkbox (- [ ])</rule>
   </validation>
 </template>

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-autopm",
-  "version": "3.23.0",
+  "version": "3.23.1",
   "description": "Autonomous Project Management Framework for Claude Code - Advanced AI-powered development automation",
   "main": "bin/autopm.js",
   "workspaces": [

package/packages/plugin-pm/plugin.json

@@ -844,6 +844,15 @@
       "description": "Provider-agnostic PM workflows"
     }
   },
+  "rules": [
+    {
+      "name": "issue-structure",
+      "file": "rules/issue-structure.xml",
+      "description": "Enforces standard issue structure with required sections",
+      "priority": "high",
+      "appliesTo": ["issue-create", "epic-decompose"]
+    }
+  ],
   "dependencies": [],
   "peerPlugins": [
     "@claudeautopm/plugin-core"

package/packages/plugin-pm/rules/issue-structure.xml

@@ -0,0 +1,56 @@
+<rule id="issue-structure" priority="high" enforcement="required">
+  <description>
+    Every GitHub issue created by ClaudeAutoPM MUST follow the standard structure.
+    Issues without proper structure waste AI tokens and lead to scope creep.
+  </description>
+
+  <template>
+    <section name="goal" required="true">One sentence: WHAT and WHY</section>
+    <section name="context" required="true">
+      WHY this issue exists — problem background, current state, dependencies, environment.
+      Must answer: what triggered this, what works/doesn't, what must exist first.
+    </section>
+    <section name="acceptance-criteria" required="true">Measurable checkboxes, not prose</section>
+    <section name="affected-files" required="true">Exact file paths with change description</section>
+    <section name="related-issues" required="true">depends on / blocks / related to #NNN, or "None"</section>
+    <section name="verification" required="true">Concrete commands to verify (curl, kubectl, npm test, etc.)</section>
+    <section name="definition-of-done" required="true">Quality gate checkboxes — tests, deploy, PR merged</section>
+    <section name="constraints" required="false">Stack, patterns, prohibitions</section>
+    <section name="out-of-scope" required="false">What NOT to touch</section>
+    <section name="suggested-approach" required="false">Brief implementation strategy</section>
+  </template>
+
+  <title-convention>
+    <format>[type][scope] Imperative description</format>
+    <types>feat, fix, refactor, chore, test, docs, ci</types>
+    <examples>
+      <example>[feat][agent] Add retry logic to ResearcherAgent</example>
+      <example>[fix][api] Handle timeout in /ingest endpoint</example>
+      <example>[refactor][db] Extract VectorStore interface</example>
+      <example>[test][plugin] Add PluginManager install scenario tests</example>
+    </examples>
+  </title-convention>
+
+  <labels>
+    <label-group name="type">feature, bug, refactor, chore</label-group>
+    <label-group name="effort">S, M, L, XL</label-group>
+    <label-group name="ai-ready">true — issue is well-structured for AI execution</label-group>
+  </labels>
+
+  <key-principles>
+    <principle>Acceptance Criteria over narrative — AI understands checkboxes better than prose</principle>
+    <principle>Explicit scope always — say what NOT to touch or AI will "improve" neighbors</principle>
+    <principle>Name files, classes, functions — AI should not guess project structure</principle>
+    <principle>Link related issues — dependencies and blockers must be visible</principle>
+  </key-principles>
+
+  <violations enforcement="auto-reject">
+    <violation>Creating issues without Context section (why it exists, current state)</violation>
+    <violation>Creating issues without Acceptance Criteria</violation>
+    <violation>Creating issues without Affected Files section</violation>
+    <violation>Creating issues without Verification Steps</violation>
+    <violation>Creating issues without Definition of Done</violation>
+    <violation>Vague goal without concrete deliverable</violation>
+    <violation>Missing Related Issues (must list links or "None")</violation>
+  </violations>
+</rule>