punt-labs/prfaq

21 stars · Last commit 2026-04-12

Amazon Working Backwards PR/FAQ process — generate well-grounded professional documents for product discovery and decision-making.

README preview

# prfaq

[![License](https://img.shields.io/github/license/punt-labs/prfaq)](LICENSE)
[![CI](https://img.shields.io/github/actions/workflow/status/punt-labs/prfaq/docs.yml?label=CI)](https://github.com/punt-labs/prfaq/actions/workflows/docs.yml)
[![Working Backwards](https://img.shields.io/badge/Working_Backwards-hypothesis-lightgrey)](./prfaq.pdf)

**Platforms:** macOS, Linux

Amazon's [Working Backwards](#what-is-working-backwards) PR/FAQ process, grounded in your data — generate, review, stress-test, and iterate on product discovery documents inside the terminal.

## What It Does

A product discovery document without evidence is fiction. `prfaq` starts from your data — customer interviews, survey results, market reports, competitive analysis, usage metrics — and builds a PR/FAQ document grounded in that evidence.

**Your research comes first.** Drop files into a `./research/` directory before running `/prfaq`. The plugin reads `.md`, `.txt`, and `.pdf` files and treats them as primary sources. Every claim in the generated document is traced back to these sources or to web research the plugin performs itself. With [punt-quarry](https://github.com/punt-labs/quarry) installed, the plugin can also search across your indexed knowledge base — any of Quarry's 30+ supported formats (PDFs, spreadsheets, presentations, source code, images, HTML, DOCX, and more).

**Three evidence sources, in priority order:**

1. **Local research files** — your `./research/` directory (highest priority)
2. **Indexed documents** — semantic search via [punt-quarry](https://github.com/punt-labs/quarry) if installed

View full repository on GitHub →