custom-agent
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes local agent files which could contain malicious instructions intended to influence the auditing or enhancement process. \n
- Ingestion points: The 'Audit Agents' (Mode 2) and 'Enhance Agent' (Mode 3) workflows read content from
.claude/agents/*.mdand~/.claude/agents/*.md. \n - Boundary markers: The instructions do not specify explicit delimiters or 'ignore' commands to isolate the content of the audited files from the agent's own instruction context. \n
- Capability inventory: The skill has the ability to read, write, and edit files, and can perform automated fixes (potentially involving command execution) upon user confirmation. \n
- Sanitization: There is no mention of sanitizing or escaping the content of the audited files before the agent processes them.
Audit Metadata