new-venture-analyst
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection through its data processing workflow.
- Ingestion points: Business assumptions (growth rates, ARPU, churn, etc.) are extracted from user prompts in SKILL.md Step 1.
- Boundary markers: None. The skill does not provide instructions to the agent to treat these inputs as untrusted or to isolate them from instructions.
- Capability inventory: Execution of the local Python script 'scripts/model_financials.py' via a shell command as described in SKILL.md.
- Sanitization: Absent at the agent level. Although the Python script uses 'argparse' with strict types (int/float), the agent lacks instructions to sanitize or escape variables before they are interpolated into the shell command string.
- [COMMAND_EXECUTION]: The skill utilizes local shell execution to ensure calculation accuracy.
- Evidence: SKILL.md Step 2 instructs the agent to execute 'python3 scripts/model_financials.py' with arguments derived from the user's input context. While this is a functional requirement, it represents an exploitable capability if combined with unsanitized inputs.
Audit Metadata