intelligems-funnel-diagnosis

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION] (LOW): The skill utilizes several shell commands (test, cp, source, python3) to verify the workspace, move script files, and execute the analysis.
  • Evidence: Step 0, 1, 2, and 4 use bash blocks to interact with the environment.
  • [CREDENTIALS_UNSAFE] (LOW): The skill handles an INTELLIGEMS_API_KEY by prompting the user for input and then writing it to a plaintext .env file.
  • Evidence: echo "INTELLIGEMS_API_KEY=<user's key>" > ~/intelligems-analytics/.env in Step 1.
  • [DATA_EXPOSURE] (INFO): The skill reads from the filesystem (specifically .env and references/) and creates directories to manage its operational state.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill processes output from an external Python script (funnel.py) to generate a conversational report. While this is the core function, there are no specific sanitization steps mentioned for the script's output before being processed by the agent.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 07:17 AM