skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/bim-consistency-checker/Gen Agent Trust Hub
bim-consistency-checker
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implementation follows secure coding practices for its intended purpose. It uses standard Python regex and logic to validate structured data provided by the user without relying on external or untrusted dependencies.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external project data. 1. Ingestion points: Project data and file paths in CSV, Excel, and JSON formats (instructions.md). 2. Boundary markers: No explicit delimiters or instructions for the agent to ignore embedded commands are present. 3. Capability inventory: The skill performs naming, parameter, and spatial checks and generates reports (SKILL.md); it lacks network operations or arbitrary command execution capabilities. 4. Sanitization: No explicit data sanitization is implemented for the report generation. The risk is assessed as safe due to the limited capability tier of the skill.
- [COMMAND_EXECUTION]: No dangerous command execution or system-level modifications were detected. The requested filesystem permissions in claw.json are consistent with the skill's stated purpose of processing and exporting BIM model data.
Audit Metadata