codingbuddy-rules 4.4.0 → 4.5.0

package/.ai-rules/adapters/claude-code.md

@@ -523,6 +523,45 @@ Each workflow mode activates different specialist agents:
 
 **Important:** Specialists from one mode do NOT carry over to the next mode. Each mode has its own recommended specialist set.
 
+### Execution Strategy Selection (MANDATORY)
+
+When `parse_mode` returns `availableStrategies`:
+
+1. **Check `availableStrategies`** in the response
+2. **If both strategies available** (`["subagent", "taskmaestro"]`), ask user with AskUserQuestion:
+   - Option A: "SubAgent (background agents, fast)" (Recommended)
+   - Option B: "TaskMaestro (tmux parallel panes, visual monitoring)"
+3. **If only `["subagent"]`** and `taskmaestroInstallHint` present:
+   - Ask: "TaskMaestro is not installed. Would you like to install it for tmux-based parallel execution?"
+   - Yes → invoke `/taskmaestro` skill to guide installation, then re-check
+   - No → proceed with subagent
+4. **Call `dispatch_agents`** with chosen `executionStrategy` parameter:
+   - `dispatch_agents({ mode, specialists, executionStrategy: "subagent" })` — existing Agent tool flow
+   - `dispatch_agents({ mode, specialists, executionStrategy: "taskmaestro" })` — returns tmux assignments
+5. **Execute** based on strategy:
+   - **subagent**: Use `dispatchParams` with Agent tool (`run_in_background: true`)
+   - **taskmaestro**: Follow `executionHint` — start panes, assign prompts, monitor, collect results
+
+### TaskMaestro Execution Flow
+
+When `executionStrategy: "taskmaestro"` is chosen, `dispatch_agents` returns:
+
+```json
+{
+  "taskmaestro": {
+    "sessionName": "eval-specialists",
+    "paneCount": 5,
+    "assignments": [
+      { "name": "security-specialist", "displayName": "Security Specialist", "prompt": "..." },
+      { "name": "performance-specialist", "displayName": "Performance Specialist", "prompt": "..." }
+    ]
+  },
+  "executionHint": "1. /taskmaestro start --panes 5\n2. ..."
+}
+```
+
+Execute by following the `executionHint` commands sequentially.
+
 ## PR All-in-One Skill
 
 Unified commit and PR workflow that:

package/.ai-rules/rules/core.md

@@ -546,7 +546,8 @@ To preserve this implementation session for future reference:
 - All implementations must follow the Primary Developer Agent's code quality checklist
 - Respond in the language specified in the agent's communication.language setting
 - Execute one step at a time, verify tests after each step
-- Stop and return to PLAN if blockers encountered
+- On recoverable errors (file not found, command failure, path issues), try alternatives immediately
+- Only return to PLAN on unrecoverable blockers (missing dependencies, architectural conflicts)
 
 **Verification:**
 - Agent name should appear as `## Agent : [Primary Developer Agent Name]` in response
@@ -563,6 +564,21 @@ To preserve this implementation session for future reference:
 
 ---
 
+### Error Recovery (ACT mode)
+
+- On recoverable errors (file not found, command failure, path issues), try alternatives immediately
+- Do not wait for user confirmation on recoverable errors — recover immediately
+- Only report to user on unrecoverable errors (auth failure, critical data loss risk)
+- Never break the workflow on error — try alternatives and continue
+
+### Path Safety (monorepo)
+
+- Always verify current directory (pwd) or use absolute paths before git add/commit
+- In monorepo subdirectories, use git -C "$REPO_ROOT" flag for git commands
+- After using cd, ensure relative paths are not double-applied in subsequent commands
+
+---
+
 ### Eval Mode
 
 **Important:**

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "codingbuddy-rules",
-  "version": "4.4.0",
+  "version": "4.5.0",
   "description": "AI coding rules for consistent practices across AI assistants",
   "main": "index.js",
   "types": "index.d.ts",