agent-self-evaluation

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content, including agent outputs and task descriptions, to generate evaluation reports and instructions for improvement.
  • Ingestion points: The scripts/evaluate.py script reads data from stdin or user-specified files via the --task and --output arguments.
  • Boundary markers: There are no explicit delimiters or protective instructions used to separate untrusted data from the scoring logic.
  • Capability inventory: The skill performs text analysis using regular expressions and prints results; it does not have network access or capabilities to execute the analyzed content as code.
  • Sanitization: The tool does not sanitize or escape the external content it evaluates.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 03:40 PM
Security Audit — agent-trust-hub — agent-self-evaluation