cst-simulation-workflow
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses specialized tools (
cst_add_to_history_toolandcst_run_python_tool) to drive the CST Studio Suite environment. These tools facilitate the execution of VBA macros and Python scripts, which are necessary for complex electromagnetic modeling and simulation tasks. The instructions emphasize using small, named chunks and avoiding ad hoc scripts where native tools exist. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data from existing CST project files, including parameter names, geometry definitions, and result tree entries. This represents a potential surface for indirect prompt injection if a project file contains malicious instructions in its metadata.
- Ingestion points: Existing
.cstproject files (references/workflow.md), parameter lists, and user-supplied simulation specifications. - Boundary markers: No explicit delimiters are specified for data ingested from existing projects.
- Capability inventory: The skill has access to history injection (VBA), Python execution, and file system operations for project management.
- Sanitization: The risk is mitigated by explicit instructions to inspect projects before modification and to work only on copied versions of user projects to prevent data loss or unauthorized changes.
Audit Metadata