The Agent Skills Directory

[SAFE]: The skill provides best-practice guidance for database security, including the correct implementation of Row-Level Security (RLS) and warnings against exposing credentials like SecretKey or API Key in client-side code.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect injection as it ingests data from database environment responses (queryEnv) and query execution plans (queryPgDatabase).
Ingestion points: Data enters the context via the output of queryPgDatabase (SQL EXPLAIN results) and queryEnv (environment info).
Boundary markers: The instructions do not define specific delimiters for separating tool output from agent instructions.
Capability inventory: The skill uses applyMigration for DDL schema changes and queryPgDatabase for SQL execution.
Sanitization: The skill explicitly notes a platform-level read-only gate that prevents the ANALYZE keyword, mitigating the risk of side-effect-based attacks during diagnostics.
[COMMAND_EXECUTION]: The skill instructs the agent on how to use management tools (applyMigration, queryPgDatabase) for database administration. This is the core functionality and is used with appropriate safety warnings.

postgresql-best-practices-cloudbase