polardbx-ttl20

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill processes user-supplied database schemas to generate SQL statements.
  • Ingestion points: User-provided CREATE TABLE statements (SKILL.md, references/ttl20-user-guide.md).
  • Boundary markers: None present.
  • Capability inventory: No tool access or shell execution capabilities are defined or used; the skill only generates text output.
  • Sanitization: None present.
  • [SAFE]: Static detections for concealment were evaluated as false positives. Phrases such as 'metadata-only change' and 'async, long-running' correctly describe the performance characteristics of PolarDB-X database operations and are not attempts to hide malicious agent behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 06:44 PM
Security Audit — agent-trust-hub — polardbx-ttl20