skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/open-construction-estimate/Gen Agent Trust Hub
open-construction-estimate
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches pre-trained embedding models from Hugging Face using the
sentence-transformerslibrary. Hugging Face is a well-known industry service for artificial intelligence assets. - [COMMAND_EXECUTION]: Utilizes standard Python libraries such as
pandasandscikit-learnfor database management and vector similarity matching operations. - [PROMPT_INJECTION]: Analyzes user-provided BIM element descriptions and properties to perform semantic matching.
- Ingestion points: element descriptions and BIM data fields in
instructions.md. - Boundary markers: Delimiters for external data are not explicitly defined in the instructions.
- Capability inventory: File system access for reading local CSV databases and writing Excel reports, plus network access for pricing APIs.
- Sanitization: Input data is processed through mathematical vectorization rather than being used in direct command execution or shell interpolation.
Audit Metadata