The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of database client binaries including sqlite3, psql, and mysql to perform data operations. It also utilizes shell features like piping to jq and I/O redirection for file creation.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions for installing necessary system dependencies using apt-get and brew, as well as standard Node.js database drivers (pg, mysql2, better-sqlite3) from the npm registry. All suggested packages are industry-standard tools for database connectivity.
[CREDENTIALS_UNSAFE]: Code examples for PostgreSQL and MySQL include placeholder credentials (e.g., PGPASSWORD=mypassword). These are used for illustrative purposes and do not represent hardcoded secrets for specific external systems.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes data from external sources (databases).
Ingestion points: Data enters the context from SQLite files, remote PostgreSQL servers, and MySQL instances.
Boundary markers: The skill lacks specific delimiters to separate database-retrieved content from system instructions, though it recommends the agent warn users about large datasets.
Capability inventory: Includes file writing (via fs.writeFileSync and shell redirection) and network-based database connections.
Sanitization: Implements basic CSV escaping for data formatting but does not include security-focused sanitization to prevent adversarial instructions within the database content from influencing the agent's logic.

database-query-and-export