diagnose-and-fix-tracked-errors

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external data such as error logs and stack traces, which creates an indirect prompt injection surface.
  • Ingestion points: Tracked error events, logs, and stack traces processed by the method in SKILL.md.
  • Boundary markers: None explicitly defined in the skill instructions.
  • Capability inventory: The agent is expected to execute tests and reproduction scripts, perform file system writes for fixes, and interact with external observability tool APIs.
  • Sanitization: No specific input sanitization is prescribed beyond the diagnostic method's emphasis on deterministic reproduction.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 07:40 PM
Security Audit — agent-trust-hub — diagnose-and-fix-tracked-errors