skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/oce-tendering/Gen Agent Trust Hub
oce-tendering
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of informational content and workflow descriptions for an ERP system. No executable scripts, command-line operations, or network exfiltration patterns were found.\n- [SAFE]: The identified API patterns use relative paths for internal ERP communication and do not point to suspicious external domains.\n- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow involving the ingestion of external bidder data (bids and RFIs). While this presents a potential indirect prompt injection surface, no exploitable implementation or malicious instructions are present in the skill itself. Ingestion points: bidder responses and bid data; Boundary markers: not specified; Capability inventory: ERP API access; Sanitization: not specified.
Audit Metadata