skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/bim-validation-pipeline/Gen Agent Trust Hub
bim-validation-pipeline
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external IFC models and XML Information Delivery Specification (IDS) files provided by the user.
- Ingestion points: The skill reads external data via
ifcopenshell.open()in theBIMValidatorclass andxml.etree.ElementTree.parse()in theIDSValidatorclass. - Capability inventory: The skill is restricted to local file system operations (reading models and writing Excel reports). It lacks network exfiltration capabilities, subprocess spawning, or arbitrary code execution paths.
- Sanitization: While naming convention checks use regular expressions, other element properties are ingested and processed without sanitization before being included in the generated reports.
- Boundary markers: The implementation does not include specific delimiters or "ignore" instructions to prevent the agent from potentially interpreting malicious data embedded in the BIM files as instructions.
Audit Metadata