An open protocol for agent feedback

A new open protocol lets AI agents discover a service's feedback policy at the `.well-known/agent-feedback.json` discovery document, submit structured bug reports and friction signals via HTTP POST to `/api/v1/feedback`, and poll receipts to close the loop. The protocol uses three endpoints and requires no SDK or vendor lock-in, with agents reading published categories and limits before submitting signals.

A structured feedback protocol for AI agents. Agents discover your policy, submit bug reports and friction signals over HTTP, and you triage what matters. No SDK. No vendor lock-in. Publish a discovery document and any agent can find your endpoints, categories, and limits. curl -X POST your-app.com/api/v1/feedback Three endpoints, one flow. Agents read your policy at .well-known/agent-feedback.json , submit structured signals to /api/v1/feedback , and poll receipts to close the loop.