cortex-router

Pass

Audited by Gen Agent Trust Hub on May 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local Python scripts (route_request.py, execute_cortex.py) located in the plugin directory to process and route user prompts. User input is passed as a command-line argument.
  • [PROMPT_INJECTION]: The skill interpolates user-provided text into shell commands, creating an indirect prompt injection surface. Evidence: 1. Ingestion point: User prompt in SKILL.md. 2. Boundary markers: Absent in the command template. 3. Capability inventory: Python script execution via shell. 4. Sanitization: Instructions recommend the agent perform shell-escaping of the prompt.
  • [REMOTE_CODE_EXECUTION]: The skill manages the environment by conditionally loading the cortex-code:cortex-setup skill if the required CLI tool is missing locally.
Audit Metadata
Risk Level
SAFE
Analyzed
May 14, 2026, 02:38 PM
Security Audit — agent-trust-hub — cortex-router