backend-design-optimizer
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION] (LOW): The skill has an Indirect Prompt Injection surface (Category 8) as it synthesizes content from external research agents. Ingestion points: Research prompts for codex-deep-research and gemini-research-analyst (SKILL.md). Boundary markers: Absent; research results are directly interpolated into recommendations. Capability inventory: Generates architecture patterns, SQL schemas, and TypeScript code for user review. Sanitization: None. Risk is low because the agent does not execute the code itself.
- [SAFE] (SAFE): No malicious patterns such as remote code execution, credential exfiltration, or persistence mechanisms were detected.
Audit Metadata