security-research

Pass

Audited by Gen Agent Trust Hub on Jun 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is designed for professional security auditing and includes strict instructions to ensure all testing is performed safely and non-destructively.\n- [PROMPT_INJECTION]: The skill interacts with untrusted codebases, creating a surface for indirect prompt injection. This risk is managed through a multi-agent verification workflow.\n
  • Ingestion points: Accesses codebase content, diffs, and logs in the working tree (Phase 0 and Phase 1).\n
  • Boundary markers: There are no explicit delimiters used to separate the content of audited files from the hunter agents' instructions.\n
  • Capability inventory: Utilizes team_create to instantiate agents with the capability to perform file searches and execute Proof of Concept code.\n
  • Sanitization: Mitigation is provided via explicit instructions to use 'toy inputs', 'local-only execution', and 'dry-run proof' methods.\n- [COMMAND_EXECUTION]: The workflow requires agents to create and run code to prove exploitability. The skill provides clear guidelines to keep these actions safe and limited to the local environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 4, 2026, 11:45 AM
Security Audit — agent-trust-hub — security-research