guidance
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill facilitates the use of the Guidance library (developed by Microsoft Research) to constrain LLM generation using regex and grammars. The instructional content and code examples are valid and promote security by enforcing structured output and preventing hallucinations. No malicious behavior was detected.- [EXTERNAL_DOWNLOADS]: Installation instructions refer to standard, reputable Python packages including 'guidance' and 'transformers'. These are well-known tools from trusted tech organizations with official support.- [COMMAND_EXECUTION]: The provided 'scripts/safe_arithmetic.py' script implements a math-only parser using the Python AST module. It includes safety checks for expression length, depth, and allowed operators, effectively preventing arbitrary code execution while providing necessary calculation functionality for the agent's reasoning workflows.
Audit Metadata