Skip to content

About — HARP

About HARP

An open protocol ensuring every AI agent action receives explicit human authorization.

HARP (Human Authorization & Review Protocol) is an open, cryptographically verifiable protocol that ensures every AI agent action receives explicit human approval before execution.

As AI coding agents become increasingly autonomous — generating plans, modifying files, running commands, and deploying code — HARP provides the missing authorization layer: a secure, out-of-band mechanism for humans to review and cryptographically sign every action.

  • Open standard — HARP is a protocol specification, not a product. Anyone can implement it.
  • Zero-knowledge — The relay infrastructure never sees plaintext code or diffs.
  • Fail-closed — If verification fails for any reason, execution is denied.
  • Interoperable — Works with any AI agent, IDE, or mobile platform.

HARP is developed in the open on GitHub. Contributions, feedback, and implementation reports are welcome.