skills/win-hao/stop-me/stop-me-check/Gen Agent Trust Hub

stop-me-check

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill audits untrusted codebases by reading configuration files like .stop-me.yml, logs in DECISIONS.md, and general source code. This represents a significant attack surface where malicious instructions embedded in these files could potentially override the agent's logic or behavior.
  • Ingestion points: The agent reads .stop-me.yml, DECISIONS.md, and all files in the repository during the audit sweep.
  • Boundary markers: The skill instructions do not specify any boundary markers or instructions for the agent to ignore potentially malicious content within the audited files.
  • Capability inventory: The skill uses file system read access and executes shell commands (git log) to analyze the repository.
  • Sanitization: No sanitization or validation of the content of the repository files is mentioned before the agent processes them.
  • [COMMAND_EXECUTION]: The skill explicitly instructs the agent to execute shell commands, such as git log -S, to search through the repository's history for secret literals. While this is the intended purpose of the auditing tool, it involves programmatic interaction with the host environment's shell based on the target repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 09:54 PM