audit-architecture
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to audit entire codebases by inspecting their files and inventorying subsystems. Since the codebase content is external and potentially untrusted, it presents a surface for indirect prompt injection where malicious instructions embedded in comments or strings within the code could attempt to influence the agent's behavior during the audit process.
- Evidence: The instructions require the agent to 'Inspect the repository and inventory every identifiable subsystem' and 'fan out bounded read-only agents per subsystem' (SKILL.md).
- Mitigation: The skill implements a highly structured output format and a strict read-only policy which significantly limits the exploitable capabilities available to any potential injection.
Audit Metadata