loop-circuit-breaker

Pass

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface by instructing the agent to carefully analyze error messages originating from potentially untrusted tools to diagnose loop root causes.
  • Ingestion points: Untrusted tool inputs and error strings are passed to the check.py script as command-line arguments.
  • Boundary markers: No delimiters or isolation instructions are provided to ensure the agent treats error content as untrusted data.
  • Capability inventory: The skill maintains its own session state through local file writes to the ~/.openclaw directory.
  • Sanitization: While tool signatures are hashed for identification, the raw error text is stored and presented back to the agent without sanitization or escaping.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 21, 2026, 05:55 AM