Quick answers to the most common questions about PM Skills.
No. Most PMs regularly use 5-10 skills. Start with the ones that match your immediate needs:
- New to PM → Start with
problem-statement,user-story, andprioritization-advisor - Experienced PM → Jump to the skills for your current challenge
- Learning PM → Work through one domain per week
See the quick decision guide:
| Your situation | Best choice |
|---|---|
| "I want the quickest start" | Paste into any AI chat |
| "I use Claude regularly" | Claude Desktop/Web |
| "I use ChatGPT regularly" | ChatGPT Projects |
| "I want the best quality" | Claude Code or Gemini CLI |
| "I want reusable team tools" | ChatGPT Custom GPTs |
Yes. PM Skills are plain Markdown files. Any AI that can read text can use them. The tool-specific guides are optimization tips, not requirements.
Yes. The entire skill library is open source. The AI tools you use with it may have their own costs.
Three ways:
- By task — Check the Getting Started table that maps "I want to..." to skills
- By domain — Browse the Catalog organized by PM lifecycle stage
- By trigger — Each skill's YAML frontmatter lists natural language triggers
| Type | Behavior | Think of it as... |
|---|---|---|
| 🧱 Component | Fill-in template → artifact | A form + checklist |
| 🔄 Interactive | Q&A → customized guidance → artifact | A consultation |
| 🎭 Workflow | Multi-phase process → multiple artifacts | A guided project |
Yes. See Skill Chaining & Workflows for patterns. The basic approach:
Complete Skill A → use its output as input for Skill B → continue to Skill C
| Score | Quality | What to do |
|---|---|---|
| 9-10 | Excellent | Share with stakeholders confidently |
| 7-8 | Good | Fix the lowest dimension, then share |
| 5-6 | Needs work | Revise 2+ dimensions |
| < 5 | Start over | Add more context, re-read the skill |
Add this instruction at the start of your conversation:
Follow the uploaded skill file's sections in strict order:
Purpose → Key Concepts → Application → Examples → Quality Rubric → Pitfalls.
Use the fill-in templates from the Application section exactly as written.
Self-score every output against the Quality Rubric.
Absolutely. Skills are Markdown files — fork the repo and customize:
- Add your product's context to examples
- Adjust quality rubric dimensions for your standards
- Add team-specific pitfalls you've encountered
- Modify templates to match your internal formats
Yes. See Building Your Own Skill for the full guide. Use the SKILL_TEMPLATE.md to get started.
You can, but it's better to use skill chaining instead. Single skills that try to do too much produce worse output than focused skills chained together.
If you cloned the repo, pull the latest changes:
git pull origin mainIf you uploaded files, re-download updated versions from the repository.
Three factors:
- Context provided — more context = better output
- AI model capability — more capable models produce better results
- Skill complexity — beginner skills are more straightforward; advanced skills need more context to shine
- Add context — product, market, team, constraints
- Use the rubric — generate → score → fix weak dimensions → regenerate
- Try a better model — GPT-4o/Claude Sonnet/Gemini Pro for complex skills
- Chain skills — use a discovery skill before a specification skill
Skills are based on widely-recognized PM frameworks (RICE, Kano, OKRs, JTBD, etc.), but:
- Frameworks are tools, not rules — adapt them to your context
- No framework fits every situation — the "When NOT to Use This" sections help avoid misapplication
- External references in each skill link to original sources for deeper learning
- Read Building Your Own Skill
- Copy
SKILL_TEMPLATE.md - Create your skill in
skills/[domain]/[skill-name]/SKILL.md - Follow the quality checklist
- Submit a PR
Look for domains with fewer skills or gaps in the Catalog. Common requests:
- Industry-specific adaptations (fintech, healthcare, marketplace)
- More workflow skills (Growth and Analytics domains have none currently)
- Advanced leadership skills (board management, M&A product integration)