product-ai-toolkit
CLAUDE.md4 starsChanged 6 months agoMITView on GitHub
You have the AI PM Kit plugin installed. This gives you PM-specific commands, skills, and agents. For complex multi-step work, prefer delegation to agents. For everything else, just do it. Workflow for complex tasks: 1. Check existing knowledge first — Use /recall to see what you already know 2. Learn if needed — Use /learn to explore and understand unfamiliar topics 3. Delegate to agents — Use the Task tool with appropriate agents for parallel or deep work 4. Use…
In this repository17 other files it gives its agents.
- churn-signalSkill
- debateSkill
- decision-briefSkill
- design-an-interfaceSkill
- docxSkill
- impact-sizingSkill
- interview-analysisSkill
- learnSkill
- parallel-prototypeSkill
- pdfSkill
- pptxSkill
- rapid-prototypeSkill
- recallSkill
- roadmap-narrativeSkill
- sprint-healthSkill
- web-browserSkill
- xlsxSkill
The file
We split this CLAUDE.md into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Anonymise customer data in outputs: When producing interview analyses, impact sizing, or any deliverable that references specific …
- Distinctive: line 30
Use the Task tool to delegate to these specialists:
- Distinctive: line 43
All commands accept flexible input — from one-liners to full documents
- Distinctive: line 44
All output is markdown, ready for Confluence, Slack, or Jira
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AI PM Kit
You have the AI PM Kit plugin installed. This gives you PM-specific commands, skills, and agents.
Smart Delegation
For complex multi-step work, prefer delegation to agents. For everything else, just do it.
Workflow for complex tasks:
- Check existing knowledge first — Use
/recallto see what you already know - Learn if needed — Use
/learnto explore and understand unfamiliar topics - Delegate to agents — Use the Task tool with appropriate agents for parallel or deep work
- Use skills — If a skill might apply (even 1% chance), invoke it
Don't delegate simple, single-step tasks, quick questions, or synthesis work.
Data Security & Privacy
Many workflows in this kit connect to production systems (Amplitude, Jira, Confluence, Slack, Glean) and process sensitive data. Follow these principles:
- Assume MCP data is sensitive. Data pulled from Amplitude, Jira, Confluence, and Slack is live production data — treat it as confidential by default.
- Anonymise customer data in outputs. When producing interview analyses, impact sizing, or any deliverable that references specific customers, anonymise names and identifying details unless the user explicitly asks for real names.
- Be careful with web search. Competitive analysis and market research use external search. Avoid including your company name, internal product names, or strategic context in search queries — these are visible to search providers.
- Knowledge files persist indefinitely.
/learnsaves to~/.claude/knowledge/with no expiry. Don't store customer-specific data, financial figures, or confidential strategy in knowledge files — they accumulate over time and aren't encrypted. - Financial data needs care. Impact sizing outputs contain customer names, deal values, ARR, and churn data. Store workbooks securely and don't share via unsecured channels.
- Review before posting. Workflows that post to Slack, create Confluence pages, or create Jira issues are visible to others. Check sensitivity before pushing to shared systems.
Available Agents
Use the Task tool to delegate to these specialists:
- market-research-analyst — Competitive intelligence, market sizing, research
- cpto-review — Executive-level review of PRDs, architecture, strategy (uses Opus)
- ux-designer — UX design and user experience
- prototype-builder — Rapid prototyping
- data-brief-analyst — Data analysis and briefs
- customer-brief-analyst — Pre-meeting intel brief on any customer account (usage, sentiment, open issues, talking points)
- voice-of-customer — Synthesise ambient customer signals into confidence-ranked insight themes for roadmap planning
- data-storyteller — Turn Amplitude metrics or raw data into a narrative insight brief with "so what" framing
Key Patterns
- All commands accept flexible input — from one-liners to full documents
- All output is markdown, ready for Confluence, Slack, or Jira
- Skills that need real data will attempt to use MCP integrations (Amplitude, Jira, Figma, etc.) if configured
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

