Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
28 lines (19 loc) · 1.77 KB

CONTRIBUTING.md

File metadata and controls

28 lines (19 loc) · 1.77 KB
Raw

Contributing

Thanks for looking at a BaseLayer agent. These are small, personally-operated tools, but PRs and issues are welcome.

Dev setup

Every repo follows the same shape:

git clone https://github.com/BaseLayerAI/<repo>
cd <repo>
cp .env.example .env   # fill in your own values — see the README's env table
  • Node repos: Node 20+, npm install, then the npm run scripts listed in package.json.
  • Python repos: Python 3.12+, install from requirements.txt or pyproject.toml.
  • Bash repos: no install; scripts are self-contained. Run bash -n on anything you touch.

Before you open a PR

  1. Pass the lint gate. CI runs the org's reusable workflow: lint/typecheck/build/test for Node, ruff + compileall for Python, shellcheck --severity=error for shell. Run the equivalent locally first.
  2. Use conventional commits. feat:, fix:, docs:, chore:, refactor:, test: — scoped where it helps (fix(collector): ...). Squash-merge is the only merge mode, so the PR title becomes the commit.
  3. Never include secrets. No tokens, keys, cookies, session state, or real recipient addresses — in code, in tests, in fixtures, or in PR descriptions and logs. Configuration goes through env vars; if you add one, update .env.example and the README's env table in the same PR. GitHub push protection is enabled, but treat it as a backstop, not the gate.
  4. Update the docs you invalidated. If behavior, schedules, or env vars changed, the README (and docs/RUNBOOK.md where present) change in the same PR.

Scope

These agents are deliberately single-purpose. PRs that harden, simplify, or fix are an easy yes; PRs that grow an agent into a platform are probably a no — open an issue first.