The Agent Skills Directory

[DYNAMIC_EXECUTION]: The template code for database health monitoring uses dynamic loading to access library functions at runtime.
Evidence: The function check_sqlalchemy_engine_health in templates/connection_pool.py uses __import__("sqlalchemy") to dynamically load the library and execute a connectivity test.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon untrusted data from the user's codebase, such as database schemas and migration scripts, creating an attack surface for embedded instructions.
Ingestion points: SQL migration files, SQLAlchemy/Django/Prisma models, and database schema definitions provided in the agent's context.
Capability inventory: The skill provides patterns for managing database connections, executing schema migrations, and managing indexes.
Boundary markers: While the skill promotes safe database practices, it does not explicitly define prompt-level delimiters to separate instructions from the processed database code.
Sanitization: The instructions mitigate runtime risks by enforcing strict timeouts (lock_timeout, statement_timeout) and batching large operations.

production-postgres