Data Flow Diagrams
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFENO_CODE
Full Analysis
- PROMPT_INJECTION (SAFE): No attempts to override agent behavior or bypass safety filters were detected. The instructions are focused on the logical methodology of DFD creation.
- DATA_EXFILTRATION (SAFE): No hardcoded credentials, sensitive file paths, or network operations are present.
- REMOTE_CODE_EXECUTION (SAFE): The skill does not download or execute external scripts or packages.
- COMMAND_EXECUTION (SAFE): No shell commands or system-level operations are defined.
- INDIRECT_PROMPT_INJECTION (LOW): While the skill processes user-provided descriptions of systems to generate diagrams, it lacks any exploitable capabilities (such as file writing or network access) that could be triggered by malicious data. Evidence Chain: 1. Ingestion points: User-provided system descriptions in the conversational context. 2. Boundary markers: Absent. 3. Capability inventory: None (the skill is limited to text and Mermaid diagram generation). 4. Sanitization: Absent.
Audit Metadata