upgrading-to-prisma-6
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- SAFE (SAFE): No malicious patterns, prompt injections, or security vulnerabilities were detected. The skill focuses strictly on legitimate software migration tasks.\n- Indirect Prompt Injection (SAFE): 1. Ingestion points: The skill identifies and reads 'prisma/schema.prisma' and source code files (*.ts, *.js) using Grep and Glob. 2. Boundary markers: No explicit delimiters or instructions are used to separate user code from agent instructions. 3. Capability inventory: Includes file system access (Read, Write, Edit) and recommendations for command execution (npm install, npx prisma). 4. Sanitization: User code is processed without explicit sanitization. While this establishes a surface for indirect prompt injection, the risk is negligible as the operations are standard for a code transformation skill.
Audit Metadata