AI Agent Incidents
100% in your browser · data never leaves your device
A rotating feed of AI agent incidents and risks: failures, bugs, hacks, hallucinations, jailbreaks and leaks — pulled live from Hacker News and Google News. No key required.
How to use
- Just hit Run — no query needed, the feed rotates through risk topics.
- Read headlines, click through to the source.
- Use it as a daily risk scan for autonomous agents.
- Agents call the paid POST /api/ai-incidents endpoint ($0.001).
Where it applies
- Risk monitoring for an autonomous agent fleet.
- Context for trust decisions: is the ecosystem having a bad day?
- A research signal for agent incident databases.