Skills
skills/akillness/skills-template/agent-evaluation/Gen Agent Trust Hub

agent-evaluation

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill contains educational Python snippets illustrating the use of subprocess.run to execute external testing tools like pytest for grading coding tasks.\n- [REMOTE_CODE_EXECUTION]: The examples demonstrate how to evaluate agent-generated Python code using the exec() function. This is provided for instructional purposes with an explicit note to perform the execution within a sandbox environment.\n- [EXTERNAL_DOWNLOADS]: The documentation references trusted industry resources and well-known services, including Anthropic's engineering blog, SWE-bench, WebArena, and standard GitHub Actions for CI/CD integration.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 12:43 PM