tam-sam-som-calculator
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides an ingestion point for untrusted data by instructing users to paste marketing materials, website copy, and customer testimonials in "Step 0: Gather Context". This content is interpolated into the agent's context without boundary markers or sanitization, potentially allowing for prompt injection attacks hidden in the external text.
- Ingestion points:
SKILL.md(Step 0: Gather Context). - Boundary markers: None specified.
- Capability inventory: Local script execution (
scripts/market-sizing.py) and potential network operations via the referencedautonomous-investigationprotocol in Mode 3. - Sanitization: None specified.
- [COMMAND_EXECUTION]: The skill package includes a Python script
scripts/market-sizing.pyand instructions inSKILL.mdfor executing it via the command line. While the script itself is benign and lacks network or sensitive file access, it demonstrates a workflow involving shell command execution.
Audit Metadata