skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/cwicr-report-generator/Gen Agent Trust Hub
cwicr-report-generator
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external files (CSV, Excel, JSON) and interpolates it directly into HTML report templates using Python f-strings without sanitization or escaping. Ingestion points: Data enters through CSV, Excel, or JSON file parsing and is processed into DataFrames in SKILL.md. Boundary markers: The generated HTML output does not use delimiters or isolation markers to separate user-controlled data from instructions. Capability inventory: The skill possesses the capability to write files to the local file system through methods like save_html and generate_excel. Sanitization: The implementation lacks HTML escaping or string validation for fields such as item descriptions or category names, allowing raw data injection into the final report.
Audit Metadata