skill-doctor
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a specialized developer tool for static analysis and does not contain malicious code, unauthorized network access, or attempts at credential exfiltration.
- [INDIRECT_PROMPT_INJECTION]: The skill has a potential attack surface because it processes third-party instruction files, but it implements clear safety protocols.
- Ingestion points: The skill reads local SKILL.md files and their referenced resources during the audit process.
- Boundary markers: The instructions explicitly mandate treating audited resources as evidence and not as instructions governing the agent's behavior during the audit.
- Capability inventory: The skill's capabilities are limited to local file system reading and user-approved patching; it does not execute audited skills or perform network operations.
- Sanitization: All repairs are presented individually for human review and require manual approval before being applied to the file system.
Audit Metadata