The Agent Skills Directory

[SAFE]: The skill provides comprehensive and standard technical guidance for PostgreSQL database administration. All code snippets and CLI commands align with documented best practices for DBA tasks, and placeholders are consistently used for all credentials and sensitive parameters.
[COMMAND_EXECUTION]: The skill includes shell commands for database maintenance and replication (e.g., pg_basebackup, systemctl, pg_ctl) in references/replication.md and references/maintenance.md. These are standard administrative tools consistent with the skill's primary purpose and use-case.
[INDIRECT_PROMPT_INJECTION]: The skill involves processing data from database query outputs and statistics. As per the mandatory evidence chain:
Ingestion points: Data enters context through EXPLAIN ANALYZE and pg_stat view results as described in SKILL.md and references/performance.md.
Boundary markers: None explicitly defined for these data inputs.
Capability inventory: SQL command execution (CREATE, DROP, ALTER, UPDATE) and shell commands for service management (pg_ctl, systemctl).
Sanitization: None specified for tool output processing. The risk is assessed as safe given the technical context and the skill's function as a specialist reference guide.

db-postgres