testing-tool-evaluator
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and evaluate untrusted external data such as vendor claims and tool configurations. Ingestion points: The agent is instructed in SKILL.md to gathering 'vendor claims' and 'actual user data' for evaluation. Boundary markers: Absent; there are no clear delimiters or instructions to ignore embedded commands in external data. Capability inventory: The skill includes a Python script that uses the 'requests' library to interact with external services. Sanitization: Absent; the provided code logic does not include validation for external input or URLs.- [DATA_EXFILTRATION]: The skill provides a Python framework that performs network operations to non-whitelisted domains, posing a risk of SSRF. Evidence: The '_test_performance' method in SKILL.md uses 'requests.get(api_endpoint)' where the endpoint is sourced from a potentially untrusted configuration dictionary without verification.
Audit Metadata