clickhouse
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill functions as an expert knowledge base for ClickHouse, containing 28 rules and 15 reference files. The instructions focus on improving performance, storage efficiency, and operational stability.
- [SAFE]: Hardcoded credentials found in
references/backup-restore.mdandreferences/integrations.md(e.g.,AKIAIOSFODNN7EXAMPLE,password123) are clearly identified as placeholders. The skill includes a dedicated 'Security Considerations' section inSKILL.mdadvising users to never use these in production and recommending the use of environment variables, secret managers, or Kubernetes secrets. - [SAFE]: External downloads and dependencies mentioned (e.g.,
clickhouse-backup,clickhouse-sqlalchemy) are standard tools within the ClickHouse ecosystem. Links point to official documentation (clickhouse.com) and established community repositories on GitHub. - [SAFE]: The skill uses natural instructional language to guide the agent's behavior. The use of 'IMPORTANT' and 'MUST' headers is consistent with technical documentation and does not constitute a malicious prompt injection intended to bypass safety filters.
- [SAFE]: There is no evidence of obfuscation (Base64, zero-width characters), persistence mechanisms, or unauthorized network exfiltration. All documented commands and SQL patterns are relevant to the stated purpose of ClickHouse administration.
Audit Metadata