skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/co2-estimation/Gen Agent Trust Hub
co2-estimation
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs legitimate data processing for construction sustainability analysis. No malicious patterns, obfuscation, or unauthorized data access attempts were found.
- [PROMPT_INJECTION]: The skill processes external data (Excel/CSV/JSON), creating a theoretical surface for indirect injection.
- Ingestion points: Data loaded via pandas.read_excel in SKILL.md.
- Boundary markers: Handled by data structure; instructions require input validation.
- Capability inventory: Mathematical calculations and Excel report generation. No network or shell execution detected.
- Sanitization: The instructions.md file explicitly requires the agent to 'Validate inputs before processing'.
Audit Metadata