Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
49 lines (38 loc) · 1.88 KB

CONTRIBUTING.md

File metadata and controls

49 lines (38 loc) · 1.88 KB
Raw

Contributing

This project uses DCO signoff instead of a CLA.

Signed-off-by: Your Name <you@example.com>

Good first contributions:

  • Add deterministic BountyBench fixtures.
  • Add bounty templates with verifiable acceptance criteria.
  • Add verifier plugins and fixture tests.
  • Improve MCP/OpenAPI agent ergonomics.
  • Add GitHub dogfooding bounties.

Every issue, bounty, and PR should answer two distribution questions when possible:

  • How did you find Agent Bounties?
  • What made this bounty or project worth participating in?
  • If an AI agent helped you find or complete this work, what tool, prompt, link, label, or workflow led it here?

Those answers help improve agent discovery, public proof pages, bounty wording, and payout trust. They are not part of bounty acceptance or payout approval. Maintainers should ask once on the relevant issue or PR when a participant has not answered yet; see docs/distribution-learning.md for the prompt and current adoption signals. Maintainers can aggregate answered comments with cargo run -p cli -- discovery-report --input-fixture crates\cli\fixtures\discovery_answers.json and use the output to improve bounty labels, proof pages, MCP discovery, and payment-trust messaging.

If your PR is useful but not ready for main, maintainers may copy it to a collab/pr-<number>-<topic> branch so other contributors can target follow-up PRs there. That keeps iteration moving while protected payment, workflow, contract, and release paths stay out of main until they pass the normal gates. Collaboration branches are not bounty acceptance or payout approval.

Before running expensive checks, run the preflight:

.\scripts\preflight.ps1 -Mode core
.\scripts\preflight.ps1 -Mode full

Real-money code paths require tests for idempotency, replay safety, and audit links from request to proof to settlement.