The Agent Skills Directory

[SAFE]: The skill primarily consists of ClickHouse SQL queries designed to audit system state. All operations are read-only (SELECT) against system tables like system.dictionaries, system.asynchronous_metrics, and system.events.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves data from system.text_log and system.query_log, which are populated by activity on the ClickHouse server.
Ingestion points: checks.sql contains queries targeting the message column in system.text_log and the query column in system.query_log.
Boundary markers: The skill does not use specific delimiters or instructions to the agent to ignore potentially malicious content within these log entries.
Capability inventory: The skill is restricted to standard SQL query execution within the database environment.
Sanitization: No sanitization is performed on the log data before it is presented to the agent. This represents a minor surface for indirect injection via database logs, though it is considered acceptable for the primary purpose of database diagnostics.

altinity-expert-clickhouse-dictionaries