agent-router
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use
pnpm execandmise execto run local management tools. It states that theai-cliwrapper bypasses provider permissions, highlighting the need for isolation. - [INDIRECT_PROMPT_INJECTION]: The skill's primary function of reviewing external code checkouts and processing remote prompts creates an injection surface.
- Ingestion points: Untrusted data is ingested from the
workFolder,prompt-file, and code checkouts. - Boundary markers: No specific delimiters or instructions to ignore embedded content are defined.
- Capability inventory: The skill can execute local CLI commands and route data to external model providers.
- Sanitization: Content is not sanitized; security relies on the operational boundary of an isolated worktree.
Audit Metadata