The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data such as database schemas, migration scripts, and configuration files during review and audit workflows (SKILL.md, references/migration-patterns.md).
Ingestion points: The skill processes SQL scripts, ORM configurations, and database engine settings (e.g., pg_hba.conf, mongod.conf) provided by the user or found in the local codebase.
Boundary markers: The skill includes an explicit "AI Self-Check" and specific "Rules" in SKILL.md to ensure the agent does not adopt unsafe practices or ignore safety filters when processing data.
Capability inventory: The skill leverages shell-based database clients (psql, mysql, mongosh, sqlcmd) and system utilities for verification, and it can generate audit reports in the file system.
Sanitization: Instructions mandate the use of parameterized queries and strict data validation to mitigate the risk of processing malicious inputs.
[EXTERNAL_DOWNLOADS]: The skill references official documentation and well-known administrative tools from established technology providers.
Fetches technical guidance from official sources for PostgreSQL, Redis, and Valkey (references/redis-valkey.md).
References widely-used industry tools for migrations and audits, including pgloader, Percona Toolkit, and gh-ost (references/migration-patterns.md).
[COMMAND_EXECUTION]: The skill provides diagnostic and administrative commands for database maintenance and operational workflows (SKILL.md, references/backup-patterns.md).
Recommends the use of database client utilities to verify connectivity, configuration settings, and query plans.
Employs standard system utilities such as tar, zstd, gpg, and age for performing backups, compression, and encryption of database data.

databases