The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external databases, which serves as an entry point for untrusted content that could influence agent behavior. * Ingestion points: Data retrieved from the database via db.orm.<Model>.all(), db.sql.<Table>.all(), and runtime.execute(plan) as described in SKILL.md, postgres.md, and mongo.md. * Boundary markers: No specific delimiters or 'ignore embedded instructions' warnings are established for the data returned from database queries. * Capability inventory: The skill provides the agent with full CRUD capabilities (create, read, update, delete) and the ability to execute complex database query plans. * Sanitization: The instructions do not define sanitization, validation, or filtering procedures for data retrieved from external sources before it is processed by the agent.

prisma-next-queries