postgresql-performance

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious instructions, hidden code, or dangerous command execution patterns were identified in the skill files.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides templates for analyzing query execution plans and database performance statistics, which are common diagnostic data sources.
  • Ingestion points: SQL queries and performance metrics fetched from the database or provided by the user (SKILL.md).
  • Boundary markers: None identified, as the skill serves as a static reference for optimization patterns.
  • Capability inventory: The included validation script (scripts/validate.py) is restricted to local file path checks and contains no network or broad file system capabilities.
  • Sanitization: The skill employs yaml.safe_load() for configuration processing, mitigating risks associated with untrusted data deserialization.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 05:59 PM
Security Audit — agent-trust-hub — postgresql-performance