entity-model-auditor
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill provides a structured framework for auditing entity models by applying specific classification logic to user-provided documents and schemas.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from PRDs and codebase files (DB models, API types, schemas), which is a standard surface for indirect prompt injection. However, the skill only performs analysis and generates markdown output without invoking dangerous tools, network operations, or file system writes, rendering this surface negligible in risk.
- Ingestion points: User-provided PRD and codebase files in Step 1 of
SKILL.md. - Boundary markers: None explicitly mentioned in prompt construction.
- Capability inventory: None; the skill is restricted to analysis and text generation.
- Sanitization: Not required for this execution flow.
Audit Metadata