matlab
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied MATLAB source code, HDF5/MAT metadata, and JSON project manifests (ingestion points in scan_m_code.py and validate_project_manifest.py). While processing external content creates an attack surface, the risk is mitigated by explicit boundary markers and warnings in the documentation and tool outputs. The skill's capabilities are limited to generating scaffolds and command plans, and it implements robust sanitization including strict path resolution, symlink rejection, and identifier validation.
- [EXTERNAL_DOWNLOADS]: Documentation and helper scripts reference official repositories and registries such as PyPI and the MathWorks documentation site for version pinning and configuration guidance. These references involve well-known technology organizations and do not involve the execution of untrusted remote content.
- [COMMAND_EXECUTION]: The skill includes a utility to generate command-line plans for MATLAB and Octave execution. It incorporates safety checks for identifiers and paths, and it does not invoke the shell or run the generated commands itself, requiring explicit external execution by the user.
Audit Metadata