clickhouse-core-workflow-a
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides templates for programmatically applying Data Definition Language (DDL) statements to a ClickHouse database. This creates a vulnerability surface if the agent generates and executes these statements based on untrusted user requirements without performing strict validation.
- Ingestion points: User-provided specifications for table schemas, including column names, types, and engine parameters.
- Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions embedded within the user-provided data.
- Capability inventory: The skill is configured to allow
Bashexecution and provides templates for the@clickhouse/clientlibrary, which can perform administrative database operations. - Sanitization: The provided Node.js implementation guide lacks input sanitization or parameterization logic for the DDL strings.
- [EXTERNAL_DOWNLOADS]: The skill instructions and documentation reference the
@clickhouse/clientNode.js package, which is the official client library for ClickHouse.
Audit Metadata