The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides implementation patterns for data ingestion and database persistence, creating an indirect prompt injection surface.
Ingestion points: The UserCRUD class in SKILL.md defines methods such as create, update, and get_by_email that accept and process external input data.
Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to disregard potential instructions embedded within the data retrieved from or sent to the database.
Capability inventory: The skill provides full database lifecycle management capabilities, including reading, writing, updating, and deleting records via AsyncSession operations in SKILL.md.
Sanitization: The implementation examples do not demonstrate input validation, escaping, or sanitization protocols for external data before it is interpolated into database queries or persisted.

sqlalchemy