원클릭으로
pm-improvement-self-improvement
PM self-improvement patterns and reflection practices
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
PM self-improvement patterns and reflection practices
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | pm-improvement-self-improvement |
| description | PM self-improvement patterns and reflection practices |
| category | improvement |
"The PM must model continuous learning - you cannot orchestrate what you don't practice."
Use during skill_research phase to improve PM's own capabilities.
| Area | Skills to Improve | Triggers |
|---|---|---|
| Task Selection | pm-organization-task-selection, prioritization algorithms | Tasks skipped, wrong priority order |
| PRD Management | pm-organization-prd-reorganization, lifecycle management | Completed tasks not moved, backlog not refilled |
| Coordination | shared-worktree, parallel assignment | Build conflicts between agents |
| Communication | Agent communication protocols, message routing | Missed messages, delayed responses |
After each task cycle, reflect on these areas:
**Self-Assessment Questions**:
- Did I select the highest priority unblocked task?
- Were all dependencies verified before assignment?
- Did I check for parallel assignment opportunities?
- Was the task moved from backlog to active queue properly?
**Improvement Actions**:
- Update task selection algorithm in pm-organization-task-selection
- Add new prioritization rules based on learnings
- Document edge cases encountered
**Self-Assessment Questions**:
- Were completed tasks moved to prd_completed.json promptly?
- Was the backlog refilled when items dropped below 5?
- Are the stats (completed, pending, backlog) accurate?
- Did I maintain the 3-file lifecycle correctly?
**Improvement Actions**:
- Update cleanup procedures in pm-workflow
- Add PRD lifecycle automation
- Improve stats tracking accuracy
**Self-Assessment Questions**:
- Did I detect when both Developer and Tech Artist were idle?
- Did I check for file conflicts before parallel assignment?
- Were status updates sent correctly to watchdog?
- Did I prevent worktree conflicts?
**Improvement Actions**:
- Update conflict detection in pm-organization-task-selection
- Improve parallel assignment checks
- Document safe parallel task combinations
**Self-Assessment Questions**:
- Were my messages clear and actionable?
- Did I use proper message format and IDs?
- Were status updates sent at appropriate times?
- Did I respond to questions from agents promptly?
**Improvement Actions**:
- Update message templates in shared-core
- Improve response timeliness
- Document communication patterns
❌ DON'T:
✅ DO:
| Trigger | Action |
|---|---|
| Task failed after 3 retries | Analyze why, improve task spec |
| QA found bugs in "clean" code | Review code review patterns |
| Agent asked for clarification | Update relevant skill |
| PRD reorganization needed | Improve prd-reorganization patterns |
| Worktree conflict occurred | Improve conflict detection |
| Missed parallel opportunity | Add parallel check to task selection |
After each retrospective, verify:
Research from feat-027 (Star Rating Preview):
Complete Developer workflow orchestration - task research sequence, implementation flow, validation gates, PRD synchronization, exit conditions.
Complete Game Designer workflow - skill invocation protocol, GDD creation, playtest flow with GDD review, design sessions. MUST load before starting assignments.
Complete PM Coordinator workflow - task assignment, project orchestration, PRD management, worker coordination. Use proactively when starting PM agent work.
Complete PM Coordinator workflow - task assignment, project orchestration, PRD management, worker coordination. Use proactively when starting PM agent work.
Complete QA Validator workflow orchestration. References specialized skills for each validation step. Load at session startup for full protocol.
Base instructions and guidelines for all agents in the system. This skill provides foundational behaviors and communication protocols that all agents should follow.