pre-launch-data-check
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted CSV data which could potentially contain malicious instructions targeting the AI agent. However, the script's logic is confined to data validation and the agent follows a rigid reporting template, significantly mitigating the risk.
- Ingestion points: The
scripts/check_list.pyscript reads user-provided leads CSV files and suppression files via thecsvmodule. - Boundary markers: The script's output does not include explicit boundary markers for data values, but the agent's instructions provide a structured 'Output format' to separate data interpretation from command execution.
- Capability inventory: The skill uses a Python script limited to standard library modules. It lacks network capabilities and does not execute external shell commands. File operations are limited to reading input CSVs and writing a list of errors to a specified output file.
- Sanitization: While the script uses a standard CSV parser to prevent structural injection, it does not explicitly sanitize the text content of the CSV fields before printing them in the audit report.
Audit Metadata