eval-analyze
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes instructions to gather and process experiment results, metrics, and segments from the environment or user input. This represents a surface for indirect prompt injection, as the agent is directed to analyze potentially untrusted data.
- Ingestion points: Step 1 (Gather Context) specifically instructs the agent to collect control/treatment metrics and sample sizes from the context.
- Boundary markers: The skill lacks explicit delimiters or instructions to treat gathered data as untrusted content or to ignore embedded instructions.
- Capability inventory: According to the frontmatter, the agent has access to
Bash,Write, andWebFetchtools. - Sanitization: No specific data validation or sanitization steps are defined for the input experiment results.
- [SAFE]: No hardcoded credentials, malicious remote code execution patterns, or persistence mechanisms were detected. The use of standard tools for data analysis is consistent with the skill's stated purpose.
Audit Metadata