freecad-scripts
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documents the use of
FreeCADGui.doCommand()andrunpy.run_path()for programmatically running macros and commands within the FreeCAD environment. These are standard API features for CAD automation and include instructional warnings about executing only trusted code. - [DATA_EXFILTRATION]: The skill provides examples for exporting 3D models and saving screenshots to the local file system using methods like
Part.export(),Mesh.write(), andview.saveImage(). These operations are fundamental to the skill's purpose of CAD automation and use placeholder local paths. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to translate natural language modeling descriptions into Python scripts with file-writing and command-execution capabilities, which presents a surface for indirect prompt injection.
- Ingestion points: Natural language task descriptions provided by the user (as described in
SKILL.md). - Boundary markers: No specific delimiters or warnings for the agent to ignore instructions embedded in user modeling requests are provided.
- Capability inventory: The skill documents extensive file writing (exports), GUI manipulation, and command execution (
FreeCADGui.doCommand,runpy) through the FreeCAD API across all reference files. - Sanitization: No input sanitization or validation logic is provided in the reference templates or script generation instructions.
Audit Metadata