skill-creator

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run and subprocess.Popen in several scripts (scripts/run_eval.py, scripts/improve_description.py, eval-viewer/generate_review.py) to execute local benchmarking scripts and the claude CLI. These operations are essential for its function as an evaluation and optimization tool and operate on user-defined skill directories and test cases.
  • [EXTERNAL_DOWNLOADS]: The eval-viewer/viewer.html file includes a reference to a well-known library, SheetJS (cdn.sheetjs.com), which is used solely for rendering Excel file outputs in the browser-based review tool. This is a common practice for frontend data visualization and does not involve remote code execution within the agent environment.
  • [SAFE]: The skill instructions in SKILL.md follow best practices, including guidance for the AI agent to avoid creating malicious skills (Principle of Lack of Surprise) and implementing progressive disclosure for context management.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 04:50 AM