Thank you for your interest in ProofPath.
ProofPath is an early-stage open-source project exploring verifiable intent and causal authorization for HTTPS APIs and AI-agent actions.
We are building a small, practical layer that helps systems answer:
- Why did this action happen?
- Which intent authorized it?
- Which prior decision is its causal parent?
- Is the action reversible, compensatable, or irreversible?
- Can the action be audited after the fact?
- Improve examples in the HTTP profile.
- Add valid and invalid conformance fixtures.
- Expand the threat model.
- Implement simple Rust parser types.
- Add CLI output examples.
- Improve documentation clarity.
- Do not replace HTTPS.
- Do not replace OAuth/OIDC.
- Keep v0.1 minimal.
- Prefer explicit reason codes over vague errors.
- Make every critical action auditable.
- Treat irreversible actions as special.
Use short, clear commit messages:
Add ProofPath HTTP profile
Add initial threat model
Implement header parser
Add invalid missing-intent fixture
A useful PR should include:
- What changed.
- Why it matters.
- Any open questions.
- Tests or examples when applicable.
ProofPath discusses security and accountability. Keep discussions precise, respectful, and evidence-driven.