generation-router
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input to classify and route media generation requests. 1. Ingestion points: User descriptions of media generation or analysis tasks (SKILL.md). 2. Boundary markers: The skill is directed to produce a compact, structured route artifact. 3. Capability inventory: Capabilities include routing to other skills and executing vendor-specific CLI commands. 4. Sanitization: The skill instructions do not specify any validation or sanitization for the data interpreted from user input.
- [COMMAND_EXECUTION]: The skill includes logic to execute vendor-specific CLI commands, particularly 'postplus quote confirm', to handle transaction challenges. The instruction to 'run it directly' based on matching user input to commands creates a surface where malformed input could attempt to influence command execution paths.
Audit Metadata