Skills
Navigate
skills/matlab/simulink-agentic-toolkit/simulink-author-classic-autosar-swc/Gen Agent Trust Hub

simulink-author-classic-autosar-swc

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external ARXML files and reads generated C source files, creating an attack surface where maliciously crafted data could attempt to influence agent behavior.
  • Ingestion points: Usage of arxml.importer and fileread in references/variants-and-arxml-roundtrip.md and references/verify-generated-artifacts.md.
  • Boundary markers: SKILL.md includes explicit instructions to wait for user confirmation before performing destructive operations like autosar.api.create(modelName, "init").
  • Capability inventory: The skill utilizes evaluate_matlab_code for API calls and model_edit for Simulink model modifications.
  • Sanitization: Content is processed via official Simulink and AUTOSAR Blockset APIs.
  • [SAFE]: The skill is authored by the product vendor and uses legitimate, documented APIs and required-tools for its stated purpose.
  • [SAFE]: The instructions emphasize model integrity and safety by prioritizing incremental updates and requiring explicit authorization for resets.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 04:35 AM