clickhouse-io
Pass
Audited by Gen Agent Trust Hub on Mar 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill uses environment variables (e.g., CLICKHOUSE_PASSWORD, DATABASE_URL) to handle sensitive credentials, which is an industry-standard security practice.- [SAFE]: The provided code snippets use well-known and reputable libraries such as the clickhouse client and the pg (node-postgres) driver.- [SAFE]: The skill includes an ingestion surface where data is retrieved from PostgreSQL or streams and inserted into ClickHouse. While this pattern represents a potential surface for indirect prompt injection, it is presented as a standard architectural pattern for ETL/CDC workflows without malicious intent.
- Ingestion points: Data retrieved via extractFromPostgres() and PostgreSQL notification payloads in SKILL.md.
- Boundary markers: Not explicitly defined in the documentation snippets.
- Capability inventory: Data manipulation via clickhouse.query and clickhouse.insert functions.
- Sanitization: The examples demonstrate batch insertion logic; developers using these patterns should ensure input sanitization in their final implementations.
Audit Metadata