Documented AI coding agent failures — severity-scored, source-verified, CC BY 4.0. Callable over MCP.
-
Updated
Aug 18, 2026
#
incident-database
Here are 4 public repositories matching this topic...
PipeRoll - Agent Incident Registry: verified public records of AI-agent failures. Open data, open submissions, verified registration.
-
Updated
Sep 9, 2026 - Python
Machine-readable preflight for software agents: pass the next unsent tool call; receive ≤2 revision-pinned failure cards or typed UNKNOWN. Offline/zero-install; never authorizes execution.
offline-first mcp zero-install security-advisories ai-agents preflight machine-readable llm-security function-calling agent-tools agentic-ai tool-calling ai-agent-security agent-safety agent-guardrails incident-database devops-safety destructive-operations
-
Updated
Aug 16, 2026 - JavaScript
The ARE Incident Database: an OWASP-ASI-indexed registry of real agent failures, with honest coverage boundaries.
owasp dataset ai-safety ai-agents ai-security prompt-injection llm-security agentic-security agent-reliability-engineering incident-database
-
Updated
Aug 27, 2026 - Python
Add this topic to your repo
To associate your repository with the incident-database topic, visit your repo's landing page and select "manage topics."