Contributing

Issues and focused pull requests are welcome. The full guides live in the repository:

Reporting bugs

Open a GitHub Issue with: reproduction steps, expected vs actual behavior, and your environment (Node version, OS, relevant .knowflowrc settings with secrets removed).

Pull requests

Fork and branch (git checkout -b feature/amazing)
Conventional Commits (feat:, fix:, docs: …)
Make sure npm run check and npm test pass

Code style: shell scripts use set -euo pipe fail discipline, Node code is ES Modules, Python is 3.10+ with type annotations.

Contributing wiki pages

You can contribute knowledge, not just code — run knowflow ingest on a great URL, then have your coding agent synthesize it following the synthesis contract. Synthesized pages land as status: pending, ready for review.

A quick map for AI assistants

If you are an AI assistant working on this repository, this table is for you:

Task	File
Understand the architecture	Architecture
Understand the methodology	Methodology
Understand the data formats	Data model
Synthesize wiki pages from raw material	`docs/agents/synthesize.md`
Change extraction logic	`scripts/ingest.sh`
Change templates	`templates/*.md`
Change CLI commands	`bin/knowflow.js`
Change graph building	`scripts/graph_builder.py`
Change vector search	`scripts/vector-store.mjs`

KnowFlow is a knowledge compiler, not a search tool. The value chain:

raw material → [any agent, one synthesis contract] → pending wiki → [your review] → linked wiki → [relations] → graph + optional vector index