prd-driven-ddd
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions define a purely conceptual workflow for architectural design and modeling. No evidence of malicious code, obfuscation, or unauthorized system access was found.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface through PRDs, flowcharts, and prototypes. Evidence: (1) Ingestion points: Product requirements documents and flowcharts defined in SKILL.md. (2) Boundary markers: None explicitly mentioned. (3) Capability inventory: Generation of text-based domain models and Mermaid diagrams. (4) Sanitization: None. This surface is inherent to the skill's primary purpose and lacks functional triggers for exploitation, posing no security threat in this context.
Audit Metadata