🧠 PM Skills — 207 Professional Agent Skills for Claude, ChatGPT, Gemini, Cursor, Codex & Hermes

Generic AI gives you filler. These give you the structure a senior pro actually uses — PRDs, exec updates, launch plans, postmortems — as open-source SKILL.md files. Across 21 professions, not just product management. One source, every AI tool.

🆕 Now plugs into your stack — automate skills in n8n, build apps on Lovable, and run them in your Obsidian vault, all via a read-only REST API on the hosted Worker.

⭐ If this saves you time, star the repo — it's the #1 way to help others find it.

PM stands for Professional, not just Product Management. 207 professional skills + 4 agent templates across 28 bundles covering 21 professions. Built for Claude Code — and now portable to ChatGPT, Gemini, and Hermes Agent. Built by a PM, used by everyone.

A community-built library of professional skills for every field — product management, engineering, customer success, marketing, social media, writers, design, legal, finance, HR, sales, operations, research, and more. Each skill is a structured SKILL.md file that teaches an AI assistant how to produce professional-grade outputs for your workflows. Skills run natively in Claude Code and Hermes Agent (same open SKILL.md standard), and ship as ready-to-paste exports for ChatGPT and Gemini — see Works With.

🆕 Latest release (v29.0.0 — Multilingual + a polished experience): run any skill in 10 languages (Spanish, Mandarin, Hindi, Arabic — with RTL — Portuguese, French, German, Japanese, Russian, Indonesian); the frameworks are universal, so the model localizes the whole output. Plus a 15-page Practical Guide and a one-page cheatsheet (PDF/PNG), a navigation + playground UX revamp (a repo link from the playground, sticky nav, a redesigned home), and quality hardening — case studies, opt-in 👍/👎 feedback, and a CI regression gate that blocks score drops. Earlier (v28.4 — run it free, no credit card): the playground no longer needs a paid key to start. Use a free Google AI Studio key (no card, ~30 seconds) or pick In-browser to run a model with no key at all (WebLLM on WebGPU — fully private, nothing leaves your device), and every skill now shows a pre-generated sample output so you can see the value with zero setup. Plus the Professional Brain is now in the browser — a 🧠 Brain page that feeds your durable memory into every run and saves outcomes back. Earlier (v28.3 — the Professional Brain): skills gained a durable, local markdown memory they read before answering, write back to (provenance-tagged, approval-gated), and even act on (open tickets/post updates via action-runner, dry-run + approval). Set it up in 5 minutes — Quickstart · architecture. Recent integrations: n8n, Lovable, Obsidian + a read-only REST API (/v1) on the hosted Worker. Earlier (v28.0): every channel is now shipped & public — pip install pm-skills (LangChain/CrewAI), the hosted MCP connector https://pm-skills-mcp.pm-claude-skills.workers.dev/ (add to ChatGPT/Claude.ai/Cursor by URL), and the official MCP registry listing. See the new 🌍 Use It Anywhere section. Also: multi-model playground (run any skill with your own Claude / OpenAI / Gemini / Ollama key — local models included), a remote MCP connector (add PM Skills to ChatGPT, Claude.ai & Cursor with one URL — no install), the MCP server now exposes skills as prompts + resources, 4 new IDE targets (Cline, Continue, Zed, Roo — 9 platforms total), a Python package (pip install pm-skills → LangChain/CrewAI tools), an llms.txt for agent discovery, a reusable score-skills GitHub Action, and MCP-registry manifests. Plus the full library is now eval-scored: 196 skills, avg 4.8/5. 207 skills. See the changelog.

▶ See it in action — try the live Skill Playground

_{👆 Pick any skill, fill a short form, and run it with your own key — no install required.}

👋 New here? Start in 30 seconds

The problem: Ask any AI for a PRD, an exec update, or a launch plan and you get generic — plausible-sounding filler you still have to rewrite from scratch. The model doesn't know what "good" actually looks like for professional work.

What this fixes: Each skill is a battle-tested SKILL.md that teaches the AI the real structure, rigour, and judgement a senior professional uses — so the first draft is one you can ship, not one you have to redo.

Try these 5 first — no install needed, run them right in the live Playground:

→ Want proof first? See real sample outputs from each skill. Like what you see? Install in 2 minutes · browse all 207 skills · ⭐ star the repo so others find it.

📄 One-page cheatsheet

The whole library on one page — install, all 207 skills by profession, the Professional Brain, the REST API, integrations, and quality tooling. Great to print, share, or drop into a slide.

📥 Markdown · 🖼️ PNG · 📄 PDF — regenerate the image/PDF with node web/docs-assets/shoot.mjs (PAGES=cheatsheet).

📖 Want the full walkthrough? The Practical Guide is a 15-page, example-driven manual — install → your first skill → worked examples (PRD, exec update, RICE, postmortem) → chaining recipes → the Brain → running anywhere → tips. Download the PDF →

🧠 The Professional Brain — local-first memory for any AI agent

Generic AI forgets everything between sessions, so you re-paste context forever and last quarter's why evaporates. The Professional Brain is the fix — and it's the most novel piece here, worth a look on its own:

• Plain markdown, no vector DB — a brain/ folder (knowledge · decisions · hypotheses · stakeholders) your AI reads before answering and writes to after. Grep-able, auditable, Obsidian-compatible — memory you can read and correct by hand, not a black box.
• Provenance over confidence theatre — every fact is tagged [data] [interview] [external] [verbal] [hunch], so a hunch never poses as a measured result.
• It acts, safely — the action-runner turns a skill's recommendations into real tickets/messages: dry-run, risk-rated, approval-gated, then recorded back. Nothing acts silently.

→ Read the architecture · 5-min Quickstart · try it in the browser

🔄 One library, the whole professional workflow

These 207 skills aren't a random catalog — they cover the full arc of professional work, end to end. Wherever you are in the loop, there's a skill for it:

flowchart LR
    A(["🔍 Discover"]) --> B(["🎯 Decide"]) --> C(["🔨 Build"]) --> D(["🚀 Ship"]) --> E(["📊 Measure"]) --> F(["📣 Communicate"])
    F -. "feeds the next discovery" .-> A
    classDef s fill:#1d222b,stroke:#d97757,color:#e7ebf0;
    class A,B,C,D,E,F s;

New here? Start with the top-tier skills, or jump straight to all 207 skills grouped by profession.

🧩 Workflow Recipes — chain skills into one flow

Individual skills are great. Chaining them is the superpower. A recipe runs several skills in sequence and passes each output forward as context — so a fuzzy idea comes out the other end as a finished, joined-up set of artifacts. No other skills library chains across professions like this.

/ship-a-feature  "a referral program for B2B users"

  ambiguity-resolver → prd-template → rice-prioritisation → roadmap-narrative → go-to-market
   frame the problem    spec it        prioritise it        place on roadmap     launch plan
        └──────────────── each stage's output feeds the next ────────────────┘

→ Full detail and how to add your own in WORKFLOWS.md. Recipes run as slash commands in Claude Code, or over MCP via the get_workflow tool.

Got a coding question? Just ❓ Ask — an error, a regex, a git mess, a dependency conflict, "what does this code do?" — and the right developer skill answers instantly (like StackOverflow, but the answer's already written), with a one-click path to the community Q&A for a human follow-up.

Or build your own visually. The Workflow Canvas lets you drag any skills into a custom chain and run it in the browser — each step's output feeds the next. Like n8n, but for professional thinking. Don't know which skills to use? The ✨ Auto-Agent takes a plain-English goal, plans which skills to chain, and runs them for you — each step feeding the next:

🧠 Skill Memory — set your context once, every skill uses it

Generic output is the #1 complaint with AI. Skill Memory fixes it: tell the skills who you are once, and every skill and recipe produces output already tuned to your product, audience, and voice — no re-typing.

• In Claude Code: run /setup-context (or copy templates/pm-context.example.md → pm-context.md). Skills read it as standing context.
• In the Playground: fill the 🧠 Your context box — it's saved in your browser and prepended to every run.

Without context:  "write an exec update" → generic, you rewrite it
With context:      "write an exec update" → your voice, your metrics,
                    your audience — shippable on the first try

🤖 Run a skill from a GitHub comment (ChatOps)

This repo ships a Skill Bot: comment on any issue or PR and a skill runs and replies inline.

/skill executive-update
Audience: CEO. Period: Q2. Shipped onboarding redesign; activation up; hiring behind plan.

The bot runs the skill and posts the result as a reply. /skill list shows usage. It's gated to repo collaborators (so random commenters can't trigger paid API calls) and needs an ANTHROPIC_API_KEY secret. Copy the workflow into your own repo to give your whole team skill-powered ChatOps.

✅ Eval-verified quality — not just quantity

Grounded in canonical frameworks. These aren't invented prompts — each skill encodes a proven method and cites it: RICE (Intercom), Jobs-to-be-Done (Christensen), Continuous Discovery (Teresa Torres), Porter's Five Forces, the Pyramid Principle (Minto), Google SRE, WCAG, Obviously Awesome (April Dunford), and more. The source shows as a "📚 Based on" line on every skill page and in the Playground.

And measured, not just claimed. An eval harness runs each skill against a held-out test case, then an LLM judge (Opus 4.8) rates the output on four dimensions — structure, completeness, usefulness, grounding — averaged across two models. 15 skills are eval-scored today (and climbing); the rest are reviewed against the authoring standard. A regression gate then blocks any PR that drops a skill's score, so quality can't quietly rot as models change.

The qualitative half — real time saved. Scores measure quality; case studies measure outcomes. See the kind of before → after a skill delivers (PRD: ~2 days → ~25 min), and add your own 2-minute ROI story — we'll feature it.

The loop actually catches bad skills. A recent run flagged three skills scoring ~2/5 because they asked for missing inputs instead of delivering. We added a "work from a brief" rule, re-ran, and they jumped to 4.75/5:

That's not luck — it's a self-improving pipeline (/improve) that critiques and rewrites a skill, keeping the change only if the score goes up. See the 🏆 leaderboard for live scores; run it yourself with node evals/run-evals.mjs.

See the difference for yourself. The Playground's Compare toggle runs the same inputs with and without the skill, side by side — structured, shippable output on the left; generic mush on the right:

Already have a draft? Flip on 📝 Critique mode in the Playground — or use the standalone Grade your work tool — to paste an existing PRD/roadmap/update and get a rubric score, ranked gaps, and a redline graded against the skill's framework. Before you ship, run /red-team to stress-test the plan against a room of hostile expert personas.

This whole thing is an open, reproducible benchmark for AI professional work — and it's eval-gated: every contributed skill is auto-checked for structure and scored on the rubric. The easiest way to add one is Skill Studio — describe a skill in your browser, generate a compliant SKILL.md, and open a pull request in one click (or use the form →).

⚖️ How it compares

Most skill repos are a folder of prompts. This one is a system — measured, composable, and usable in the browser:

Contents

• 👋 New here? Start in 30 seconds
• 🔄 One library, the whole professional workflow
• 🧩 Workflow Recipes — chain skills into one flow
• 🧠 Skill Memory — set your context once
• ✅ Eval-verified quality
• 🚀 Quick Install
• 🔌 Works With — Cross-Tool Compatibility
• 🤖 Subagents & Slash Commands
• 🌐 Skill Playground — try any skill in your browser
• 📦 Plugin Directory
• 🤖 Building Blocks for Agent Templates
• 🏷️ Skill Tiers — start with the strongest
• 🗂️ All 207 Skills
• 📋 Changelog
• 🤝 Contributing
• 🔗 Related Projects

🚀 Quick Install (2 minutes)

With the skills CLI (the open agent-skills installer that works across Claude Code, Cursor, Codex, OpenCode & 60+ agents) — pick from all 207 interactively:

npx skills add mohitagw15856/pm-claude-skills            # browse & pick (auto-detects your agent)
npx skills add mohitagw15856/pm-claude-skills --list     # just preview the catalog
npx skills add mohitagw15856/pm-claude-skills --skill prd-template   # grab one
npx skills add mohitagw15856/pm-claude-skills --skill '*'            # install all 207

Or our own installer — via the pm-claude-skills npm package (Windows/macOS/Linux, needs Node), which also installs subagents, slash commands & cross-tool exports:

npx pm-claude-skills add --agent claude     # or: codex · cursor · hermes · openclaw

Or one-line MCP — make all 207 skills + 5 workflow recipes available in every session of any MCP client (Claude Code, Claude Desktop, Cursor, Windsurf), no per-file install:

claude mcp add pm-skills -- npx -y pm-claude-skills-mcp

Your assistant can then "search the skills for churn" or "run the ship-a-feature workflow" on demand. Details: mcp/README.md.

In Claude Cowork — open the plugin browser → Add from GitHub → enter mohitagw15856/pm-claude-skills, then add the bundles you want. No CLI needed.

In Claude Code, run:

/plugin marketplace add mohitagw15856/pm-claude-skills

Or install by profession:

claude plugin install pm-essentials@pm-claude-skills # Core PM + Word tracked changes

claude plugin install pm-delivery@pm-claude-skills # Delivery + PowerPoint auditor

claude plugin install pm-engineering@pm-claude-skills # Engineering (35 skills) 🆕

claude plugin install pm-cs@pm-claude-skills # Customer Success 🆕

claude plugin install pm-data@pm-claude-skills # Data + chart data extractor

claude plugin install pm-legal@pm-claude-skills # Legal

claude plugin install pm-finance@pm-claude-skills # Finance

claude plugin install pm-hr@pm-claude-skills # HR

claude plugin install pm-sales@pm-claude-skills # Sales

claude plugin install pm-operations@pm-claude-skills # Operations

claude plugin install pm-research@pm-claude-skills # Research & Healthcare

claude plugin install pm-cross@pm-claude-skills # Cross-profession

claude plugin install pm-figma@pm-claude-skills # Figma

claude plugin install pm-social@pm-claude-skills # Social Media 🆕

claude plugin install pm-writers@pm-claude-skills # Writers & Content Creators 🆕

Or clone and symlink for auto-updates:

git clone https://github.com/mohitagw15856/pm-claude-skills.git ~/pm-claude-skills mkdir -p ~/.claude/skills ln -s ~/pm-claude-skills/skills/* ~/.claude/skills/

🌍 Use It Anywhere — the AI Ecosystem

The same 207 skills reach you through every channel — pick whatever fits your stack:

196 of 207 skills are eval-scored, averaging 4.8/5 — see the leaderboard.

🧠 Give your skills memory: the Professional Brain is a local markdown folder skills read, write back to, and act on. It's a folder + one file to start — 5-minute Quickstart.

🔌 Works With — Cross-Tool Compatibility

These skills were built for Claude Code, but they aren't locked to it. Each SKILL.md is two portable parts: a small frontmatter block (name + description) and a markdown body that is just a well-structured set of instructions and output templates. The body is plain English — so it works anywhere a capable model reads instructions.

There are two kinds of support. Native SKILL.md agents read the file as-is and auto-discover skills from the description frontmatter. Other tools take the markdown body as a system prompt — for those we ship ready-made exports.

In your editor (VS Code / Cursor): the vscode-extension/ brings all 207 skills into the Command Palette — search and insert a skill as context for Copilot/Cursor chat, copy it, or open it in the Playground.

What's verified vs. what varies: the skill bodies — the frameworks, rubrics, and output templates that do the actual work — are model-agnostic and have been used across Claude and other chat LLMs. Native SKILL.md agents (Claude Code, Hermes) also get the convenience layer: automatic skill discovery from the description. On chat LLMs you paste the body in manually and lose only the auto-triggering, not the substance.

Ready-to-use exports

You don't have to strip frontmatter by hand. Every skill is exported to platform-ready files under exports/, generated from the same source so nothing is maintained twice:

• ChatGPT — copy any exports/chatgpt/<bundle>/<skill>/SYSTEM_PROMPT.md straight into a Custom GPT's instructions.
• Google Gemini — copy any exports/gemini/<bundle>/<skill>/GEM_INSTRUCTIONS.md into a Gem's instructions.
• Cursor (.mdc) · Windsurf (.md) · Aider (.md) — generated rule/conventions files in exports/cursor/, exports/windsurf/, exports/aider/ (or use the installers below).

One-command install for coding agents

Cross-platform (Windows, macOS, Linux) — recommended. One Node command installs every skill where your agent discovers them. No git, no bash:

npx pm-claude-skills add --agent claude     # skills + subagents + commands → ~/.claude/
npx pm-claude-skills add --agent codex      # OpenAI Codex (or: hermes · openclaw)
npx pm-claude-skills add --agent cursor     # .mdc rules → ./.cursor/rules
npx pm-claude-skills add --agent windsurf   # .md rules → ./.windsurf/rules
npx pm-claude-skills add --agent aider      # conventions → load with: aider --read
npx pm-claude-skills list                   # all supported agents + default paths

Add --link (symlink), --target <path>, or --dry-run to any add.

Shell one-liners (macOS / Linux / Git Bash / WSL — not PowerShell):

bash <(curl -fsSL https://raw.githubusercontent.com/mohitagw15856/pm-claude-skills/main/scripts/codex-install.sh)
bash <(curl -fsSL https://raw.githubusercontent.com/mohitagw15856/pm-claude-skills/main/scripts/openclaw-install.sh)
bash <(curl -fsSL https://raw.githubusercontent.com/mohitagw15856/pm-claude-skills/main/scripts/cursor-install.sh)

Windows / PowerShell: use the npx command above (the bash <(curl …) form is a Unix idiom and won't run in PowerShell).

Already cloned? ./scripts/install.sh --agent <name> (or python3 scripts/sync-hermes-skills.py for Hermes) does the same from your checkout.

The skill body in skills/<name>/SKILL.md is the single source of truth. Regenerate the chat-LLM / Cursor exports (or add a new platform — it's a few lines in the PLATFORMS registry) with:

node scripts/build-exports.mjs            # regenerate all platform exports
node scripts/build-exports.mjs --check    # CI: fail if exports are stale

Prefer a hand-curated ChatGPT collection? There's also a companion Custom GPT library built from the same frameworks.

🤖 Subagents & Slash Commands

It's not just skills. The library also ships Claude Code subagents and slash commands built on top of the strongest skills, so common workflows are one delegation or one command away.

Subagents (agents/) — focused personas Claude delegates to automatically by description:

Slash commands (commands/) — run a skill on whatever you pass:

/prd · /rice · /sprint-plan · /health-scorecard · /retro · /exec-summary

Personas (output-styles/) — Claude Code output styles that change the assistant's whole voice and default skill loadout. Switch with /output-style:

Startup CTO · Growth Marketer · Solo Founder · Product Leader

Install everything for Claude Code in one go (skills + subagents + commands + personas):

npx pm-claude-skills add --agent claude   # ~/.claude/{skills,agents,commands,output-styles}

Commands whose skill ships a Python helper (RICE, sprint capacity, customer health) run it to compute results, not estimate them. To string these together, see the orchestration patterns (skill chains & multi-agent handoffs).

🧩 MCP Server — Skills on Demand

For MCP clients (Claude Desktop, Cline, …), there's a zero-dependency MCP server so your assistant searches and pulls skills on demand instead of installing 172 files. It exposes three tools — list_skills, search_skills, get_skill — over stdio (no network, nothing leaves your machine).

{
  "mcpServers": {
    "pm-claude-skills": { "command": "npx", "args": ["-y", "pm-claude-skills-mcp"] }
  }
}

Then ask: "search the skills for customer churn, then apply the best one to my account." Full setup in mcp/README.md.

🔌 Skills that act on your data. Run pm-skills alongside a data server (filesystem, GitHub, Postgres, Drive…) and a skill works on your real sources: "draft the PRD from GitHub issue #123," "run churn-analysis on exports/q2.csv," "open an issue per product-launch-checklist item." Copy-paste configs + worked recipes in connectors/.

pm-skills needs no auth. To connect GitHub, Slack, Notion, etc., you add that provider's MCP server with its own token/OAuth once in your client — then every skill can use it. (filesystem = a folder you grant; GitHub = a PAT; Slack/Notion = OAuth.)

⚙️ AI-Powered Tooling

Three ways to put the library to work beyond installing files:

🤖 Run a skill in your CI — GitHub Action. Auto-write PR descriptions, changelogs, release notes, or run a code-review checklist on every PR:

- uses: mohitagw15856/pm-claude-skills/action@main
  with:
    skill: pr-description-writer
    input: ${{ steps.diff.outputs.text }}
    api_key: ${{ secrets.ANTHROPIC_API_KEY }}

🏗️ Turn your docs into a skill — generate. Point it at a URL or file and it writes a SKILL.md that follows the authoring standard:

ANTHROPIC_API_KEY=sk-ant-… npx pm-claude-skills generate --from ./team-process.md

🏆 Skill Leaderboard — evals. An LLM-as-judge harness scores each skill across Claude models on structure, completeness, usefulness, and grounding. View the leaderboard →

Score the whole library, cheaply. npm run eval:gen-cases writes a representative input for all 207 skills (curated cases kept verbatim; the rest auto-generated), then npm run eval:all runs the full pass on Haiku for ~$2 total. The weakest scorers can then be auto-rewritten with npm run improve -- <skill> (/improve), which keeps a change only if the score goes up.

🌐 Skill Playground — Try Any Skill in Your Browser

▶ Live: mohitagw15856.github.io/pm-claude-skills · 📚 Browse the full skill catalog

Don't want to install anything yet? Run any of these skills from a zero-backend web app using your own Claude, OpenAI, or Gemini key. Pick a provider, pick a skill, fill in the auto-generated form, and the result streams back live. Your key is stored only in your browser (localStorage) and sent directly to the provider you chose — nothing touches a server we own.

Run it locally:

git clone https://github.com/mohitagw15856/pm-claude-skills.git
cd pm-claude-skills
node web/build-skills.mjs               # generate the skill index (skills.json)
cd web && python3 -m http.server 8000   # serve over HTTP (not file://)
# open http://localhost:8000 and paste a key from console.anthropic.com

It's fully static — deploy the web/ folder to GitHub Pages, Netlify, or Vercel with no environment variables. Full details in web/README.md.

📦 Plugin Directory

Not sure which plugin to install? Here's what each one covers:

…view the full README on GitHub.