The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides patterns for processing external user input via Express request objects. It mitigates this risk by mandating strict schema validation.
Ingestion points: Data enters through req.body, req.params, and req.query as defined in the routing and validation documentation.
Boundary markers: The skill enforces mandatory Zod schema validation to define strict data boundaries.
Capability inventory: Documented capabilities include file system access for configuration (fs.readFileSync), database operations via Prisma, and external communication with Sentry.
Sanitization: Zod's parse and transform methods are recommended to sanitize all incoming data before it reaches the service layer.
[COMMAND_EXECUTION]: The documentation includes examples of using a local test utility (node scripts/test-auth-route.js) to verify authenticated route behavior.
[EXTERNAL_DOWNLOADS]: Guidelines recommend integrating Sentry for observability, which involves data transmission to Sentry's well-known error tracking and performance monitoring infrastructure.

backend-dev-guidelines