data-flow
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface as it processes data from a codebase to generate diagrams. Evidence: 1. Ingestion Point: Codebase content is analyzed via the codebase explorer agent (Step 1). 2. Boundary Markers: None identified in instructions. 3. Capability Inventory: Rendering via the Mermaid Chart MCP tool (Step 3). 4. Sanitization: Structural metadata is extracted but no explicit sanitization is described. This surface is inherent to the tool's utility and the risk is low due to the constrained output format.
- [DATA_EXFILTRATION]: Structural metadata such as module names and call relationships are sent to the Mermaid Chart service. This is the primary intended function of the skill and utilizes a well-known technical service.
Audit Metadata