skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/cwicr-material-procurement/Gen Agent Trust Hub
cwicr-material-procurement
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it ingests untrusted project data from external sources like CSV, Excel, and Parquet files. An attacker could embed malicious instructions within data fields such as material descriptions.
- Ingestion points: The skill reads from 'ddc_cwicr_en.parquet' and accepts user-provided project data in various formats.
- Boundary markers: No explicit delimiters or boundary markers are implemented in the Python logic to separate data from instructions.
- Capability inventory: The skill possesses filesystem write capabilities through the use of 'pd.ExcelWriter' for report generation.
- Sanitization: There is no evidence of data sanitization or escaping performed on ingested content before it is processed or presented to the user.
Audit Metadata