Product NotesPM Skills — One-Page Cheatsheet

207 open-source AI skills + a memory + an action layer, for every profession. Each is a SKILL.md that teaches an AI the structure a senior pro actually uses — so the first draft is shippable, not filler. PM = Professional, not just Product Management.
207 skills 21 professions 196 eval-scored · avg 4.8/5 Claude · ChatGPT · Gemini · Cursor · Codex · Hermes MIT

⚡ Install — pick one

Claude Code (skills+agents+commands)npx pm-claude-skills add --agent claude
Any agent (Cursor/Codex/60+)npx skills add mohitagw15856/pm-claude-skills
MCP, on demandclaude mcp add pm-skills -- npx -y pm-claude-skills-mcp
Hosted MCP / connector URLpm-skills-mcp.pm-claude-skills.workers.dev
Python (LangChain/CrewAI)pip install pm-skills
Browser, your own keyopen the Playground (no install)

🧩 Compose

  • Recipes (chain skills): /ship-a-feature /close-the-quarter /launch-a-product /rescue-an-account /run-discovery /repurpose
  • Subagents (5): pm-partner · sprint-master · cs-guardian · launch-captain
  • Commands (16): /prd /rice /sprint-plan /retro /exec-summary /brain
  • Personas (5): Startup CTO · Growth Marketer · Solo Founder · Product Leader
  • Visual: Workflow Canvas (drag→chain) · Auto-Agent (goal→planned chain)

🗂️ 207 skills across 21 professions

Profession#Try first
🛠️ Product Mgmt37/prd /rice
👩‍💻 Engineering43incident-postmortem
📣 Marketing & GTM8go-to-market
📊 Data & Analytics12metric-tree-builder
🤝 Customer Success6cs-health-scorecard
⚙️ Operations10sop-writer
🖼️ Figma10figma-design-review
🌐 Cross-profession10meeting-notes · professional-brain
⚖️ Legal · 💰 Finance7·5contract-review · investor-pitch-deck
🚀 Founders · 🎓 Educators6·6startup-idea-validator · lesson-plan
🎬 Creators · ✍️ Writers6·6hook-writer · aeo-optimizer
👥 HR · 🤝 Sales5·6job-description-writer · sales-battlecard
🧑‍💼 People · 🎨 Design5·4executive-update · design-critique
📱 Social · 🏢 Strategy · 🏥 Research5·3·4social-media-strategy · literature-review
🟢 50 Production-Ready🔵 145 Stable🟡 12 Experimental

🧠 The Professional Brain — memory + actions

A local, plain-markdown memory skills read before answering and write to after. No vector DB — grep-able, auditable, Obsidian-compatible.
brain/ context.md · knowledge/ · decisions/ hypotheses/ · stakeholders/ · entities/ · source/
  • Provenance tags: [data] [interview] [external] [verbal] [hunch] — a hunch never poses as data.
  • Ops: init · ingest · recall · record · review (/brain)
  • Recall: brain_query.py — deterministic, provenance-ranked
  • Write-back: brain_write.py — append-only, dry-run, approval-gated
  • Actions (action-runner): recommendations → real tickets/messages, risk-rated 🟢🟡🔴, approved per action. Nothing acts silently.
recall → run a skill → produce → propose actions → approve → execute → record back → weekly review

🌍 Run anywhere — REST API & integrations

Read-only, no auth, CORS-open — pm-skills-mcp.pm-claude-skills.workers.dev
GET /v1/skillsall skills (?bundle ?q ?limit)
GET /v1/skills/{name}one skill (?format=md)
GET /v1/search?q=keyword search
GET /v1/workflowsrecipe chains
  • n8n — MCP Client / HTTP node into your AI node
  • Obsidian — every skill as a vault note / AI-plugin prompt
  • Lovable — skill-powered apps; a knowledge snippet for its generator
  • Exports (10): ChatGPT · Gemini · Cursor · Windsurf · Aider · Cline · Continue · Zed · Roo · Obsidian
  • 🌐 10 languages: run any skill in Spanish · Mandarin · Hindi · Arabic (RTL) · Portuguese · French · German · Japanese · Russian · Indonesian — pick one in the playground.

✅ Quality & 🛠️ tooling

  • Eval-scored, not just counted — LLM judge on structure/completeness/usefulness/grounding → 196 scored, 4.8/5
  • Self-improving/improve keeps a rewrite only if the score rises
  • PR gates — SkillCheck (structure) + Security Auditor (injection/exfiltration)
  • Explore/try: Skill Galaxy (map) · Ask (dev Q&A) · Grade your draft · Compare mode · Skill Studio (make→PR)
  • In CI: a GitHub Action runs a skill on a PR; Skill Bot runs /skill from a comment