clickhouse-io
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection via insecure database operations.
- Ingestion points: Data enters through the
bulkInsertTradesfunction (Trade objects) and the PostgreSQL notification handler (pgClient.on) in SKILL.md. - Boundary markers: Absent; there are no delimiters or instructions to treat ingested data as untrusted.
- Capability inventory: The skill performs database writes and queries using the
clickhouseandpgclients. - Sanitization: Absent; the code directly interpolates variables (e.g.,
${trade.id}) into SQL strings, which is a classic vector for injection if the source data is attacker-controlled.
Audit Metadata