skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/progress-monitoring-cv/Gen Agent Trust Hub
progress-monitoring-cv
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external data such as site photos, BIM models, and project files, creating an attack surface where malicious instructions could be embedded in data sources. * Ingestion points: Processes external images (site_photo.jpg, drone imagery), BIM renderings, and structured project data (CSV, Excel, JSON) as defined in SKILL.md and instructions.md. * Boundary markers: There are no explicit boundary markers or instructions to ignore potential commands embedded in the ingested data. * Capability inventory: The skill utilizes filesystem permissions to read input files and write Excel reports, and it performs complex image processing and machine learning inference using torch and cv2. * Sanitization: No sanitization or validation logic is implemented for the content of the ingested data files.
- [EXTERNAL_DOWNLOADS]: The skill performs downloads of external resources during model initialization. * Evidence: The Python code in SKILL.md initializes models with pretrained=True (ResNet50 and Faster R-CNN), which triggers the download of pre-trained weights from official PyTorch repositories.
Audit Metadata