pm · superclaude-org — Claude Code skill: run projects on PDCA cycles
Tracks work through Plan-Do-Check-Act, scores confidence before implementation, and restores context from your last session at startup.
npx skills add SuperClaude-Org/SuperClaude_Framework --skill pm -a claude-codeWhat pm does
The pm skill turns the agent into a project manager. On session start it looks for docs/memory/, TASK.md and KNOWLEDGE.md, then reports what happened last time, where things stand, what comes next and what is blocking.
Work follows PDCA. Plan sets the hypothesis, success criteria and risks. Do tracks tasks in TodoWrite and records errors alongside their fixes. Check asks what worked and what failed. Act writes the lesson down: a reusable pattern on success, a mistake plus prevention on failure, both into the project knowledge base.
Before touching code, the skill scores confidence on five dimensions — duplicate implementations, architecture compliance, official docs, OSS references, root cause. Above 90% it proceeds, 70–89% means alternatives first, below 70% it stops to gather more. Worth knowing: on a low score the agent halts the task instead of writing code anyway.
What it helps with
- pick up context after a break between sessions
- break a large task into steps with success criteria
- catch a duplicate implementation before it lands
- record a mistake together with its prevention
- build a reusable pattern library for the project
Example prompts for the agent
- plan the CSV export feature with a PDCA cycle
- run a confidence check before touching the auth module
- restore context from the last session and show what is left
How to install pm
1Skills CLI (Vercel)
npx skills add SuperClaude-Org/SuperClaude_Framework --skill pm -a claude-code2Manually: put the files in place
~/.claude/skills/pm/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/pm/
1Skills CLI (Vercel)
npx skills add SuperClaude-Org/SuperClaude_Framework --skill pm -a codex2Manually: put the files in place
.agents/skills/pm/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add SuperClaude-Org/SuperClaude_Framework --skill pm -a cursor2Manually: put the files in place
.cursor/skills/pm/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install SuperClaude-Org/SuperClaude_Framework pm2Manually: put the files in place
.github/skills/pm/SKILL.md1Skills CLI (Vercel)
npx skills add SuperClaude-Org/SuperClaude_Framework --skill pm -a gemini-cli2Platform CLI
gemini skills install https://github.com/SuperClaude-Org/SuperClaude_Framework --path plugins/superclaude/skills/pmManually: put the files in place
.devin/skills/pm/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add SuperClaude-Org/SuperClaude_Framework --skill pm -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about pm
How do I install pm in Claude?
Run npx skills add SuperClaude-Org/SuperClaude_Framework --skill pm -a claude-code. The skill file lands at ~/.claude/skills/pm/SKILL.md.
What does the pm skill do?
It manages development through PDCA cycles, tracks tasks and persists context across sessions. At startup it picks up the previous status and reports blockers.
What is the confidence check in pm?
A five-point assessment before implementation: duplicates, architecture compliance, official docs, OSS references, root cause. Below 70% the skill stops and asks for more information.
Source text
SKILL.md · MIT
--- name: pm description: Project management with PDCA cycles, confidence checks, and context persistence. Auto-activates at session start to restore context. Use for task planning, progress tracking, and structured development. --- # PM Agent Mode You are the Project Management Agent. Manage development through PDCA cycles. ## Session Start Protocol 1. Check for existing context (docs/memory/, TASK.md, KNOWLEDGE.md) 2. Report status to user: - Previous: [last session summary] - Progress: [current status] - Next: [planned actions] - Blockers: [issues] ## PDCA Cycle ### Plan (Hypothesis) - Define what to implement and why - Set success criteria - Identify risks ### Do (Experiment) - Track tasks with TodoWrite - Record trial-and-error, errors, solutions - Checkpoint progress regularly ### Check (Evaluation) - "What went well? What failed?" - Assess against success criteria - Identify lessons learned ### Act (Improvement) - Success: Document pattern for reuse - Failure: Document mistake with prevention measures - Update project knowledge base ## Confidence Check (before implementation) Assess confidence on 5 dimensions: 1. No duplicate implementations? (25%) 2. Architecture compliant? (25%) 3. Official docs verified? (20%) 4. OSS references checked? (15%) 5. Root cause identified? (15%) - >=90%: Proceed immediately - 70-89%: Present alternatives, investigate more - <70%: STOP and gather more information Apply this to: $ARGUMENTS
© SuperClaude-Org · MIT · SuperClaude-Org/SuperClaude_Framework
Similar skills
All skills for Agent setup: memory, context, orchestration →pm by other authors comes first.
