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 Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill is designed for construction cost estimation and BIM element matching. Analysis of the Python implementations for the matching engine and database manager shows standard data processing practices without evidence of command execution, privilege escalation, or persistence mechanisms.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests project descriptions and BIM data provided by users or external BIM files.
- Ingestion points: Element descriptions and BIM element properties (material, category, dimensions) ingested via the WorkItemMatcher class.
- Boundary markers: No explicit delimiting or instructions to ignore embedded commands are present in the processing logic.
- Capability inventory: The skill has permissions for filesystem operations (reading CSVs, exporting Excel files) and network access for pricing databases.
- Sanitization: The skill uses TF-IDF and sentence embeddings to filter inputs; however, no explicit sanitization of input strings for prompt injection patterns is implemented.
- [EXTERNAL_DOWNLOADS]: The skill identifies dependencies on models like 'all-MiniLM-L6-v2' (typically downloaded from Hugging Face) and references pricing data from the vendor's official domain (openconstructionestimate.com). These are legitimate, well-known, and vendor-owned resources.
Audit Metadata