drizzle-orm
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFEPROMPT_INJECTIONREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill enables an agent to build and execute SQL queries based on user instructions, creating a surface for indirect prompt injection where untrusted data could influence query logic.
- Ingestion points: Natural language input translated to query builder calls like
db.select().from(table).where(...)as described throughoutapi.md. - Boundary markers: The documentation emphasizes the
sqltemplate tag which provides automatic parameterization, serving as a structural safety boundary against traditional SQL injection. - Capability inventory: The skill documentation covers full database lifecycle management, including CRUD operations, transaction handling, and schema migrations using
drizzle-kit. - Sanitization: Drizzle architecture utilizes template literals for parameterization. The skill also guides users in integrating validation libraries such as Zod and Valibot to enforce schemas on input data before database persistence.
- [DYNAMIC_EXECUTION]: The skill documents features that involve dynamic logic and runtime code optimization.
- Mentions "JIT Mappers" in
api.md, which are hardcoded functions generated and compiled at runtime to transform database result rows efficiently. - Describes
sql.rawinapi.mdfor executing unparameterized SQL fragments, which requires careful implementation to avoid security pitfalls. - Includes documentation for dynamic query building via
.$dynamic()which allows runtime modification of query objects. - [EXTERNAL_DOWNLOADS]: The installation guides in
getting-started.mdandguides.mdinstruct the user to install various database drivers and utility packages from the NPM registry. - [COMMAND_EXECUTION]: The documentation describes the use of CLI tools like
drizzle-kitfor database management tasks, includinggenerate,migrate,push, andstudiocommands.
Audit Metadata