finops-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a purely analytical procedure for FinOps tasks. It does not perform unauthorized network operations, file system modifications, or credential harvesting. The scope is explicitly limited to reading cloud metadata for report generation purposes.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data sources which represents a potential injection surface.
- Ingestion points: Cloud provider billing exports and resource configuration metadata referenced in the Inputs section.
- Boundary markers: Not explicitly defined within the instruction text.
- Capability inventory: The procedure is restricted to inventorying, calculating metrics, and flagging resources; it lacks capabilities for writing to the filesystem, executing shell commands, or making external network requests.
- Sanitization: The skill contains an explicit section stating that inputs are treated as read-only and are not interpolated into commands or file paths.
Audit Metadata