simulink-explain-missing-coverage

Warn

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: MEDIUMOBFUSCATIONMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [OBFUSCATION]: The skill includes five MATLAB P-code files (scripts/resolveCoverageData.p, scripts/getCoverageSummary.p, scripts/detectDeadLogicObjectives.p, scripts/markSldvDeadLogic.p, and scripts/traceObjectiveInputs.p). P-code is a pre-compiled, binary, and obfuscated format that prevents the inspection of source code logic, effectively hiding the operations performed by the skill from security auditing.\n- [METADATA_POISONING]: The manifest.yaml file lists the script dependencies with .m (plain-text MATLAB) extensions, while the actual files provided are .p (obfuscated) files. This discrepancy is deceptive and obscures the presence of obfuscated code in the skill's distribution.\n- [DYNAMIC_EXECUTION]: The skill relies on the evaluate_matlab_code tool to run the logic contained within the obfuscated P-code files. This enables the execution of non-verifiable code within the agent's environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill contains a vulnerability surface for indirect prompt injection as it processes external, untrusted data artifacts.\n
  • Ingestion points: External coverage results are loaded via resolveCoverageData.p (e.g., .cvt files or Simulink Test result sets).\n
  • Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are visible in the skill's workflow instructions.\n
  • Capability inventory: The skill has broad execution capabilities through evaluate_matlab_code, which can perform file system and environment operations.\n
  • Sanitization: Due to the obfuscated nature of the processing scripts, it is impossible to verify if any sanitization or validation is performed on the ingested data.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 30, 2026, 06:02 PM
Security Audit — agent-trust-hub — simulink-explain-missing-coverage