altinity-expert-clickhouse-overview
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE] (SAFE): The skill accesses database metadata and system logs (e.g.,
system.metrics,system.parts,system.processes). While this data describes the internal state of the database, it is handled within the agent's context for reporting purposes only, with no evidence of external transmission or hardcoded credentials. - [INDIRECT_PROMPT_INJECTION] (SAFE): The skill reads from
system.errorsandsystem.query_log. These sources contain error messages and query text that can be influenced by external database users. However, in the context of a health overview skill, the risk is minimal and the agent's primary task is structured summary reporting. - [COMMAND_EXECUTION] (SAFE): The skill contains pre-defined SQL queries. There is no evidence of dynamic SQL construction from untrusted user input or the use of dangerous ClickHouse functions like
url(),s3(), orexecutabledictionaries that could be used for remote code execution or data exfiltration.
Audit Metadata