The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external, potentially untrusted database schemas and SQL queries to provide optimization feedback, creating a surface for indirect prompt injection.
Ingestion points: User-provided queries, schema definitions, and EXPLAIN output mentioned across SKILL.md and references/performance.md.
Boundary markers: None explicitly defined in instructions to separate data from instructions.
Capability inventory: Generates SQL commands and maintenance shell scripts (e.g., pg_basebackup, pg_repack, systemctl) as documented in references/replication.md and references/maintenance.md.
Sanitization: Explicitly instructs the agent and user to use prepared statements to mitigate SQL injection risk in the 'Constraints' section of SKILL.md.
[COMMAND_EXECUTION]: The skill instructs the agent to guide users through system-level service management and database utility execution.
Evidence: Instructions in references/replication.md include stopping and starting the PostgreSQL service via systemctl and executing backup/restore operations via pg_basebackup. These are standard administrative actions for the stated DBA role.
[EXTERNAL_DOWNLOADS]: Recommends the installation and use of standard PostgreSQL ecosystem tools.
Evidence: references/replication.md and references/maintenance.md recommend tools such as Patroni for high availability, pg_repack for online maintenance, and pg_auto_failover for automated cluster management. These are established industry-standard utilities.

postgres-pro