| name | product-expert |
| description | Build expert knowledge bases from product management data — scan issues through a product lens, extract beliefs |
| argument-hint | [init|scan|explore|topics|propose-beliefs|accept-beliefs|summary|status] |
| allowed-tools | Bash(product-expert *), Bash(uv run product-expert *), Bash(uvx *ftl-product-expert*), Read, Grep, Glob |
Product Expert
Build expert knowledge bases from product management data (GitHub, GitLab, Jira) by analyzing issues through a product lens.
How to Run
Try these in order until one works:
product-expert $ARGUMENTS (if installed via uv tool install)
uv run product-expert $ARGUMENTS (if in the repo with pyproject.toml)
uvx --from git+https://github.com/benthomasson/ftl-product-expert product-expert $ARGUMENTS (fallback)
Typical Workflow
product-expert init github owner/repo --domain "Payment platform"
product-expert scan
product-expert scan --limit 10 --all-pages
product-expert explore
product-expert explore --pick 1,3,8
product-expert topics
product-expert propose-beliefs
product-expert accept-beliefs
product-expert summary
product-expert status
Commands
init <platform> <target> — Bootstrap knowledge base (github/gitlab/jira)
scan [--limit N] [--labels L] [--page P] [--all-pages] [--jql Q] — Fetch and analyze issues with product lens
explore [--skip] [--pick N[,N,...]] [--loop N] — Work through topic queue
topics [--all] — Show exploration queue
propose-beliefs [--batch-size N] — Extract beliefs from entries
accept-beliefs [--file F] — Import accepted beliefs
summary — Synthesize product summary from beliefs
status — Dashboard
Natural Language
If the user says:
- "analyze this product" →
product-expert init github owner/repo && product-expert scan
- "what's the product state" →
product-expert summary
- "what should I look at next" →
product-expert explore
- "what are users asking for" →
product-expert scan --labels feature-request
- "extract what we've learned" →
product-expert propose-beliefs
- "how far along are we" →
product-expert status