drizzle-safe-migrations
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute standard development commands such as
bun run db:generate,bun run lint,bun run typecheck, andbun run test. These commands are standard for projects using the Bun runtime and Drizzle ORM to manage database schemas and ensure code quality within the local environment. - [INDIRECT_PROMPT_INJECTION]: The skill involves processing and verifying external data such as database migration files and journal metadata.
- Ingestion points: The agent reads generated SQL migration files and the Drizzle journal file (
drizzle/meta/_journal.json) to verify ordering and content. - Boundary markers: The skill provides clear verification checklists, decision trees, and workflow steps that act as delimiters for the agent's logic.
- Capability inventory: The skill uses subprocess calls via the
bun runcommand to interact with the project's local tooling. - Sanitization: While no automated sanitization is described, the workflow emphasizes manual human-in-the-loop verification of all generated SQL and journal artifacts before deployment.
Audit Metadata