The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external data from database outputs, including execution plans, server logs, and activity statistics.
Ingestion points: Untrusted data enters the agent's context through query results from views like pg_stat_activity and pg_stat_statements as referenced in references/05-monitoring.md.
Boundary markers: The instructions do not explicitly define delimiters for database output, which is common in technical documentation but technically increases the surface for indirect injection if a database contains malicious strings.
Capability inventory: The skill enables the agent to perform complex SQL operations including EXPLAIN ANALYZE, ALTER TABLE, and configuration changes.
Sanitization: There are no explicit sanitization routines defined for processing database results.
[COMMAND_EXECUTION]: The reference files contain a variety of SQL commands intended for database maintenance and performance tuning.
Evidence: Files such as references/01-overview.md and references/04-patroni-i-automatyczny-failover.md provide scripts for managing indices, gathering statistics, and monitoring replication.
Context: These operations are restricted to the database environment and are standard for the declared role of a PostgreSQL DBA debugging tool.

postgres-operations-debugging