speccrew 0.6.30 → 0.6.32

package/.speccrew/agents/speccrew-feature-designer.md

@@ -195,8 +195,10 @@ If resuming from an interrupted session:
 | Checkpoint | If Passed | Resume Point |
 |------------|-----------|--------------|
 | `function_decomposition.passed == true` | Skip Checkpoint A | Start from Step 3 (Frontend Design) |
-| `feature_spec_review.passed == true` | Skip Checkpoint A & B | Start from Phase 4 (API Contract) |
-| `api_contract_joint.passed == true` | All checkpoints complete | Ask user: "Feature Design phase already completed. Do you want to redo?" |
+| `feature_spec_review.passed == true` | Skip Checkpoint A & B | Start from Phase 4 (API Contract Generation) |
+| `api_contract_joint.passed == true` | Feature Spec + API Contract joint confirmation already passed | Ask user: "Feature Design phase already completed. Do you want to redo?" |
+
+> Evaluate checkpoints from most complete to least complete. If `api_contract_joint.passed == true`, it takes precedence over all earlier checkpoints.
 
 3. **Display Resume Summary**:
    ```
@@ -647,19 +649,75 @@ Invoke API Contract skill directly:
 
 ### 4.3 Multiple Feature Specs (Parallel Worker Dispatch)
 
-Invoke `speccrew-task-worker` agents in parallel:
-- Each worker receives:
-  - `skill_path`: `speccrew-fd-api-contract/SKILL.md`
-  - `context`:
-    - `feature_spec_path`: Path to one Feature Spec document
-    - `feature_id`: Feature ID (e.g., `F-CRM-01`)
-    - `feature_type`: `Page+API` or `API-only`
-    - `output_path`: `{iterations_dir}/{iteration}/02.feature-design/{feature_id}-{feature-name-slug}-api-contract.md`
+If **2+ Feature Specs** in registry:
+
+1. **Initialize DISPATCH-PROGRESS.json for API Contract stage**:
+
+   > ⚠️ Use `--tasks-file` instead of `--tasks` to avoid PowerShell JSON parsing issues.
 
-- **Parallel execution**: One worker per Feature Spec document
-- **Output file naming**:
-  - Format: `{feature-id}-{feature-name-slug}-api-contract.md`
-  - Example: `F-CRM-01-customer-list-api-contract.md`
+   ```bash
+   # Step 1: Write tasks JSON to temp file inside iteration directory
+   # Create .tasks-temp.json with the task array content
+   # Step 2: Initialize with --tasks-file
+   node {update_progress_script} init --file {iterations_dir}/{iteration}/02.feature-design/DISPATCH-PROGRESS.json --stage 02_feature_design_api_contract --tasks-file {iterations_dir}/{iteration}/02.feature-design/.tasks-temp.json
+   # Step 3: Delete .tasks-temp.json after successful init
+   ```
+
+   Example `.tasks-temp.json` content:
+   ```json
+   [{"id":"F-CRM-01"},{"id":"F-CRM-02"},{"id":"F-CRM-03"}]
+   ```
+
+2. **Dispatch Workers** (batch of 6):
+   - Each worker receives:
+     - `skill_path`: `speccrew-fd-api-contract/SKILL.md`
+     - `context`:
+       - `feature_spec_path`: Path to one Feature Spec document
+       - `feature_id`: Feature ID (e.g., `F-CRM-01`)
+       - `feature_name`: Feature name — **MUST be the exact value from .checkpoints.json, used verbatim for output filename**
+       - `feature_type`: `Page+API` or `API-only`
+       - `output_path`: `{iterations_dir}/{iteration}/02.feature-design/{feature_id}-{feature-name-slug}-api-contract.md`
+
+3. **Wait for batch completion**, update progress per worker:
+   ```bash
+   node {update_progress_script} update-task --file {iterations_dir}/{iteration}/02.feature-design/DISPATCH-PROGRESS.json --task-id {feature_id} --status completed
+   ```
+
+4. **Update `.checkpoints.json`** for each completed Feature:
+   - Set `api_contract_status` = `completed`
+
+5. **Parallel execution**: One worker per Feature Spec document
+
+6. **Output file naming**:
+   - Format: `{feature-id}-{feature-name-slug}-api-contract.md`
+   - Example: `F-CRM-01-customer-list-api-contract.md`
+
+### 4.3.1 API Contract Error Handling
+
+When any API Contract worker reports failure:
+
+1. **Update status**:
+   ```bash
+   node {update_progress_script} update-task --file {iterations_dir}/{iteration}/02.feature-design/DISPATCH-PROGRESS.json --task-id {feature_id} --status failed --error "[Phase 4] {error_message}"
+   ```
+
+2. **Continue batch**: Do NOT stop entire batch for single failure. Complete remaining workers.
+
+3. **Report to user**:
+   ```
+   📊 Phase 4 (API Contract) — Batch complete: {success}/{total} succeeded, {failed} failed
+   ├── ✅ F-SYS-01, F-SYS-02, ...
+   └── ❌ F-MEMBER-02: [error description]
+   
+   Retry failed features? (yes/skip/abort)
+   ```
+
+4. **Retry strategy**: Same as Phase 3
+   - If user says "yes" → Re-dispatch failed features in next batch (Phase 4)
+   - If user says "skip" → Mark as `skipped`, continue to joint confirmation with partial results
+   - If user says "abort" → STOP workflow, report partial results
+
+5. **Batch failure threshold**: If >50% workers in a batch fail → STOP entire workflow
 
 ### 4.4 Joint Confirmation
 
@@ -668,6 +726,10 @@ After both Feature Spec and API Contract documents are ready, present summary to
 - List all API Contract documents with paths
 - Request user confirmation before proceeding to system design phase
 - After confirmation, API Contract becomes the read-only baseline for downstream stages
+- After confirmation, update `.checkpoints.json`:
+  ```bash
+  node {update_progress_script} write-checkpoint --file {iterations_dir}/{iteration}/02.feature-design/.checkpoints.json --stage 02_feature_design --checkpoint api_contract_joint --passed true
+  ```
 
 ### 4.5 Finalize Stage (Update Workflow Progress)
 

package/.speccrew/skills/speccrew-knowledge-bizs-dispatch-xml/SKILL.md

@@ -94,6 +94,19 @@ Stage 4: System Summary
     <field name="completed_dir" required="true" type="string" desc="Marker file output directory for Worker results (absolute path required)"/>
   </block>
 
+  <!-- ============================================================
+       Global Continuous Execution Rules
+       ============================================================ -->
+  <block type="rule" id="GLOBAL-R1" level="forbidden" desc="Continuous execution constraints — NEVER violate">
+    <field name="text">DO NOT ask user "Should I continue?" or "How would you like to proceed?" between stages</field>
+    <field name="text">DO NOT offer options like "Full execution / Sample execution / Pause" — always execute ALL stages to completion</field>
+    <field name="text">DO NOT suggest "Let me split this into batches" or "Let's do this in parts" — process ALL features sequentially</field>
+    <field name="text">DO NOT pause to list what you plan to do next — just do it</field>
+    <field name="text">DO NOT ask for confirmation before generating output files</field>
+    <field name="text">DO NOT warn about "large number of files" or "this may take a while" — proceed with generation</field>
+    <field name="text">ONLY pause at explicit &lt;event action="confirm"&gt; blocks defined in the workflow</field>
+  </block>
+
   <!-- ============================================================
        Stage 0: Platform Detection
        ============================================================ -->
@@ -140,13 +153,12 @@ Discovered frontend platform: ${frontend.name} → Tech stack: Vue/React/UniApp
       </block>
     </block>
 
-    <!-- Step 4: Merge platform list -->
-    <block type="task" id="S0-B4" action="run-script" status="pending" desc="Merge and generate platform list">
-      <field name="command">node "${ide_skills_dir}/speccrew-knowledge-bizs-dispatch/scripts/platform-detector.js"</field>
-      <field name="arg">--source</field>
-      <field name="arg">${source_path}</field>
-      <field name="arg">--output</field>
-      <field name="arg">${sync_state_bizs_dir}/platforms.json</field>
+    <!-- Step 4: Generate platform list JSON -->
+    <block type="task" id="S0-B4" action="run-script" status="pending" desc="Generate platforms.json from discovered modules and frontends">
+      <field name="command">node -e "const fs=require('fs'); const data=JSON.parse(process.argv[1]); fs.mkdirSync(process.argv[2], {recursive:true}); fs.writeFileSync(process.argv[2]+'/platforms.json', JSON.stringify(data,null,2)); console.log('platforms.json written')"</field>
+      <field name="note">The LLM constructs the platforms array from S0-L1 and S0-L2 discovery results, then writes it via this inline script. No external platform-detector.js is needed.</field>
+      <field name="arg">${platforms_json_string}</field>
+      <field name="arg">${sync_state_bizs_dir}</field>
       <field name="output" var="platforms" from="${sync_state_bizs_dir}/platforms.json"/>
     </block>
 
@@ -984,3 +996,33 @@ After all 5 stages complete, return a summary object to the caller:
   }
 }
 ```
+
+---
+
+## CONTINUOUS EXECUTION RULES
+
+This skill MUST execute all stages continuously without unnecessary interruptions.
+
+### FORBIDDEN Interruptions
+
+1. DO NOT ask user "Should I continue?" after completing a stage
+2. DO NOT suggest "Let me split this into batches" or "Let's do this in parts"
+3. DO NOT pause to list what you plan to do next — just do it
+4. DO NOT ask for confirmation before generating output files
+5. DO NOT warn about "large number of files" — proceed with generation
+6. DO NOT offer "Should I proceed with the remaining items?"
+7. DO NOT present options like "Full execution / Sample execution / Pause"
+
+### When to Pause (ONLY these cases)
+
+1. Explicit `<event action="confirm">` blocks in the workflow (e.g., Stage 0 platform confirmation)
+2. Ambiguous requirements that genuinely need clarification
+3. Unrecoverable errors that prevent further progress (failure rate > 50%)
+4. Security-sensitive operations (e.g., deleting existing files)
+
+### Batch Execution Behavior
+
+- When multiple features need processing, process ALL of them sequentially without asking
+- Use marker files (.done.json) to track progress, enabling resumption if interrupted by context limits
+- If context window is approaching limit, save progress to checkpoint and inform user how to resume
+- NEVER voluntarily stop mid-batch to ask if user wants to continue

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "speccrew",
-  "version": "0.6.30",
+  "version": "0.6.32",
   "description": "Spec-Driven Development toolkit for AI-powered IDEs",
   "author": "charlesmu99",
   "repository": {