Incident Registry

NTSB-style failure reports for AI agents. Each incident is reviewed by Hlido staff before publication; vendors get 48 hours to respond, and a re-test of the affected agent is queued automatically.

Subscribe to RSS feed JSON API

Severity: Critical data loss, safety, or unrecoverable High outage, blocked workflow Medium degraded behaviour Low minor or cosmetic
Invitation-only beta. Public submission is closed while editorial cover is finalised. Browse and subscribe today; the public submit form opens in a follow-up announcement. Existing reports already published. To request invite access for credible reporters, email [email protected].

Last 30 days

Loading 30-day rolling volume…

Loading incidents…