math-tools
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the third-party Python library
sympyvia pip, as documented in the SKILL.md file. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted mathematical expressions provided as strings through various command-line operations (e.g.,
simplify,solve,derivative). - Ingestion points: Expressions enter the system as command-line arguments in
scripts/math_calculator.pyand are passed to thesafe_parsefunction. - Boundary markers: None identified. The skill relies on SymPy's parser to interpret the strings as mathematical objects.
- Capability inventory: The script performs calculations using SymPy but does not have capabilities for network operations, arbitrary file writes, or system command execution beyond its own logic.
- Sanitization: The
safe_parsefunction uses SymPy'sparse_exprwithstandard_transformations. While this limits the input to SymPy-compatible math, the resulting strings are returned to the agent context, creating a surface where an attacker could embed text intended to influence the agent's next action.
Audit Metadata