machine-learning
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill contains documentation and valid Python code examples for common machine learning tasks using standard libraries like scikit-learn, pandas, and numpy.
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The
skill-report.jsonfile defines prompt templates that interpolate user-supplied parameters (e.g., business goals, dataset columns) into the agent's context. While this constitutes an injection surface, the risk is negligible because the skill lacks executable tools or capabilities that could be exploited by malicious input. Ingestion points: User input parameters inskill-report.jsontemplates. Boundary markers: Absent. Capability inventory: None. Sanitization: Absent. - [SAFE]: The skill includes a diagnostic manifest (
skill-report.json) containing metadata and a self-audit report. Per analysis rules, this content was evaluated as data but was not treated as an authoritative claim of safety. Independent technical review of all provided instructions and reference scripts confirms the skill's safety.
Audit Metadata