sympy

Pass

Audited by Gen Agent Trust Hub on Oct 1, 2026

Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill documents SymPy's capability to generate and execute code at runtime through lambdify, autowrap, and ufuncify. These tools allow symbolic expressions to be converted into optimized Python, C, or Fortran functions.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes functions for parsing strings into symbolic expressions (e.g., parse_expr, parse_latex). These represent a surface for indirect prompt injection if external data is processed. 1. Ingestion points: Parsing utilities in references/code-generation-printing.md. 2. Boundary markers: The documentation suggests limiting transformations and using controlled dictionaries. 3. Capability inventory: Code generation and file-writing capabilities. 4. Sanitization: A validation pattern is provided that uses regular expressions to block dangerous syntax like import and __.
  • [DYNAMIC_EXECUTION]: The skill illustrates the use of pickle for serializing mathematical objects. The documentation contextualizes this for local result saving, a standard persistence pattern.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 1, 2026, 07:50 AM
Security Audit — agent-trust-hub — sympy