altinity-expert-clickhouse-dictionaries
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill performs SQL queries against standard ClickHouse system tables (system.dictionaries, system.asynchronous_metrics, etc.). These are expected administrative operations for the stated purpose.
- [DATA_EXPOSURE] (SAFE): While the skill accesses metadata about dictionaries (configuration, names, types), it does not attempt to read the actual sensitive data contained within the dictionaries or exfiltrate it to external endpoints.
- [INDIRECT_PROMPT_INJECTION] (LOW): The skill ingests data from SQL query results (such as error messages in 'last_exception'). While this is a surface for indirect injection, the skill does not have risky downstream capabilities (like file writing or network requests) that could be exploited by malicious content in a database error message.
Audit Metadata