prisma-postgres-setup
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: All network requests are directed to official Prisma domains (api.prisma.io, console.prisma.io), and all dependencies are official vendor packages or reputable community libraries.
- [COMMAND_EXECUTION]: The skill uses shell commands for project setup, API interaction, and connectivity testing. It includes logic to delete existing projects if database limits are reached, but this is gated by interactive user selection to prevent accidental data loss.
- [EXTERNAL_DOWNLOADS]: Fetches required packages including prisma, @prisma/client, and pg from the standard NPM registry.
- [CREDENTIALS_UNSAFE]: Instructs users on the secure handling of PRISMA_SERVICE_TOKEN and DATABASE_URL. It prioritizes environment variables and provides specific steps to ensure .env files are excluded from version control using .gitignore.
- [PROMPT_INJECTION]: The skill includes a feature to generate database schemas from natural language descriptions.
- Ingestion points: User-provided data model descriptions in Step 6 (SKILL.md).
- Boundary markers: None explicitly used in the prompt instructions.
- Capability inventory: Shell command execution (npx prisma) and file system writes (SKILL.md).
- Sanitization: The risk of malicious instructions in the description is mitigated by the requirement for the agent to present the generated schema for user review and approval before execution.
Audit Metadata