environmental-and-social-safeguards
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of instructional markdown and reference materials related to international development bank safeguard standards (World Bank ESF and AfDB ISS).
- [SAFE]: No executable scripts, network-active commands (e.g., curl, wget), or hardcoded credentials were detected within the skill or its associated reference files.
- [SAFE]: Capability boundaries are clearly defined, instructing the agent to default to read-only operations and seek explicit authority for sensitive actions like interviews or certification.
- [PROMPT_INJECTION]: The skill acknowledges a surface for indirect prompt injection by processing external "ToR and client" data. However, it implements basic safety logic by requiring the agent to stop or block the workflow if required inputs or permissions are missing, mitigating potential misuse through data ingestion.
Audit Metadata