intelligence-route
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches and executes the @claude-flow/cli package from the NPM registry using npx.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted task descriptions as input for its routing logic and model prediction tools. 1. Ingestion points: The argument in the SKILL.md definition. 2. Boundary markers: No specific delimiters or instruction-isolation markers are implemented for the task description. 3. Capability inventory: The skill has access to Bash execution and internal MCP tools for model routing, prediction, and success training. 4. Sanitization: There is no evidence of sanitization or validation of the task description content before it is passed to the internal intelligence tools.
Audit Metadata