The Agent Skills Directory

[SAFE]: The skill focuses exclusively on establishing best practices for database management, query optimization, and security. It proactively addresses common pitfalls such as SQL injection and improper isolation.
[INDIRECT_PROMPT_INJECTION]: While the skill involves processing user-supplied SQL and schemas, it includes explicit instructions to mitigate injection risks, such as using driver-level parameter binding and validating dynamic identifiers.
[DYNAMIC_EXECUTION]: The skill correctly identifies that EXPLAIN ANALYZE executes queries and can have non-reversible side effects (e.g., sequence increments). It recommends using disposable databases for performance testing to prevent accidental production impact.
[PRIVILEGE_ESCALATION]: The skill provides defensive guidance regarding the search_path and security-definer functions, warning against configurations that could allow unauthorized object lookup or privilege misuse.

postgresql