a2a-proof: black-box behavioral contracts for deployed A2A agents #2071
aspix2k
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
a2a-proof runs repeatable black-box checks against a deployed A2A agent. It reads the Agent Card, sends real requests, and fails CI when observable behavior drifts.
It complements protocol tooling: TCK covers conformance, ITK covers SDK interoperability, and Inspector or workbenches help with interactive debugging. a2a-proof keeps application behavior under versioned contracts.
The demo runs locally and needs no API key. Contracts can assert text, structured data, files, task lifecycle, latency across repeated trials, and signed AP2 mandate/receipt flows. A GitHub Action runs the same contracts in CI.
I’d value feedback from teams running A2A agents: which deployed behaviors are hardest to keep stable, and where does the contract or reporting UX get in the way?
All reactions