Skills
skills/clickhouse/agent-skills/clickhouse-best-practices/Gen Agent Trust Hub

clickhouse-best-practices

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of the mcp-clickhouse Python package and the addition of an MCP server via https://mcp.clickhouse.cloud/mcp. These resources are official components provided by the vendor (ClickHouse Inc).
  • [COMMAND_EXECUTION]: The skill provides instructions for agents to use the clickhouse-client CLI and curl for database interactions. These are standard methods for the intended functionality.
  • [DATA_EXFILTRATION]: The skill follows security best practices by advising against hardcoding credentials and instead recommending environment variables or pre-configured connection profiles.
  • [PROMPT_INJECTION]: No malicious prompt injection patterns were detected. The skill contains a dedicated 'Agent Query Safety' section (Rule 4.1) that enforces resource limits like LIMIT, max_execution_time, and max_rows_to_read on all agent-generated queries.
  • [SAFE]: The skill acts as a security enhancer by providing explicit rules to prevent unbounded or inefficient database queries that could impact cluster performance.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 10:26 AM