altinity-expert-clickhouse-part-log

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected. The skill is entirely composed of SQL queries and instructions for ClickHouse database diagnostics.
  • [DATA_EXFILTRATION]: The skill reads technical metadata from database system tables such as system.part_log and system.mutations. This information is used for monitoring database health (e.g., merge rates, part counts, and mutation errors). No network operations or external data transmission logic is present.
  • [PROMPT_INJECTION]: The skill processes data from the database system tables, representing an indirect injection surface if an attacker can manipulate table names or error messages.
  • Ingestion points: system.part_log and system.mutations (referenced in checks.sql)
  • Boundary markers: Absent; queries are executed directly.
  • Capability inventory: The skill contains no shell command execution (subprocess), file-writing, or network request capabilities.
  • Sanitization: None; data is processed as diagnostic metrics for display purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 07:21 PM
Security Audit — agent-trust-hub — altinity-expert-clickhouse-part-log