altinity-expert-clickhouse-reporting
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes read-only SELECT queries against ClickHouse system tables (
system.query_log,system.processes,system.query_views_log). This is the intended behavior for a performance diagnostic tool and does not involve arbitrary command execution on the host system. - [DATA_EXFILTRATION]: While the skill reads metadata about queries (including snippets of query text and user names), it does not contain any mechanisms to transmit this data to external or unauthorized locations. Accessing system logs is necessary for the skill's primary purpose of performance tuning.
- [PROMPT_INJECTION]: The instructions and SQL templates do not contain patterns intended to bypass agent safety guidelines or override system behavior.
- [SAFE]: The skill follows standard practices for database monitoring, including time-bounding queries and applying limits to result sets. The external module references are consistent with the vendor's own ecosystem (Altinity).
Audit Metadata