matlab-migrate-settings

Pass

Audited by Gen Agent Trust Hub on Aug 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses standard system utilities such as grep, find, strings, and uname to detect MATLAB installation directories and extract settings metadata from factory libraries (.so, .dll, or .hpp files). These operations are restricted to discovery within the local environment.
  • [COMMAND_EXECUTION]: A mandatory validation step uses the evaluate_matlab_code tool or the matlab command-line interface to execute the newly generated script. This dynamic execution is used solely to verify the correctness of the migration and ensure zero runtime errors in the target release.
  • [PROMPT_INJECTION]: The skill processes untrusted data in the form of user-provided .m files (Ingestion Point: <script-path> in SKILL.md). Although explicit boundary markers are absent, the analysis instructions focus on specific structural patterns (e.g., .matlab. namespace references) and the Capability Inventory is limited to local file generation and validation, which mitigates the risk of indirect injection.
  • [SAFE]: The skill performs local file read and write operations to migrate settings. It implements a safety best practice by never overwriting the original input file, instead creating a release-specific copy (e.g., startup_R2026a.m).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 2, 2026, 12:12 PM
Security Audit — agent-trust-hub — matlab-migrate-settings