postgresql-performance-expert

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • SAFE (SAFE): No malicious patterns, command execution, or sensitive data access detected. The skill provides static instructional content and SQL/JavaScript snippets for educational purposes.
  • Indirect Prompt Injection (SAFE): While the skill is designed to interact with user-provided database symptoms and queries, it lacks any exploitable capabilities (such as file-writing or network requests), rendering the ingestion of untrusted data safe.
  • Ingestion points: User input describing SQL performance issues.
  • Boundary markers: Absent.
  • Capability inventory: None.
  • Sanitization: Not applicable as the skill only provides text-based recommendations.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:32 PM