The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data in the form of SQL queries and application code, which could be used to influence the agent's behavior.
Ingestion points: User-provided SQL statements, endpoint implementations, and UI flow descriptions are ingested during the design and review workflow.
Boundary markers: The skill does not define specific markers or delimiters to isolate user-supplied content from its own instructions.
Capability inventory: The skill utilizes queryPgDatabase for SQL execution, applyMigration for database schema updates, and queryEnv for environment metadata collection.
Sanitization: The skill documents a platform-level restriction where queryPgDatabase enforces a read-only gate that rejects potentially mutating operations like ANALYZE.
[COMMAND_EXECUTION]: The skill facilitates the execution of database-related commands through platform-specific tools.
The skill instructs the agent to perform database inspections and schema migrations using queryPgDatabase, applyMigration, and queryEnv.
These operations are documented as part of the intended vendor-provided functionality for managing the CloudBase environment.

postgresql-best-practices-cloudbase