ifc-impl-property-extraction
Pass
Audited by Gen Agent Trust Hub on Aug 9, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides legitimate architectural and engineering patterns for Industry Foundation Classes (IFC) schema navigation. It covers structural data resolution across multiple schema versions including IFC2x3, IFC4, and IFC4.3.
- [SAFE]: No obfuscation, base64 encoding, or hidden malicious content was found. The skill is entirely instructional and does not request network access, sensitive file permissions, or credential storage.
- [PROMPT_INJECTION]: The skill describes how the agent should ingest data from external IFC/STEP model files. This ingestion surface is a vector for indirect prompt injection if malicious instructions are embedded within the properties or metadata of a processed model.
- Ingestion points: IFC model files (STEP format) processed via the resolution patterns defined in
SKILL.mdandreferences/examples.md. - Boundary markers: Absent; the instructions do not recommend the use of delimiters or specific safety markers to isolate extracted data from the agent's core instructions.
- Capability inventory: The skill defines read-only traversal logic for the IFC graph. No write capabilities or external network calls are present in the provided skill files.
- Sanitization: Absent; the logic focus is on schema correctness and does not include mechanisms for filtering or sanitizing the content of property values before they are integrated into the agent's context.
Audit Metadata