postgres-mcp

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (LOW): The skill possesses an attack surface for indirect prompt injection due to its SQL execution capabilities. Ingestion points: The sql parameter in sub-skills/query.md and sub-skills/analyze_query_performance.md accepts raw strings which may contain embedded instructions. Boundary markers: Absent. There are no instructions defining delimiters or clear warnings to the agent to ignore instructions embedded within the SQL data. Capability inventory: The skill enables full database query execution, schema inspection, and performance analysis. Sanitization: Absent. The skill lacks instructions for sanitizing, escaping, or validating SQL input before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:31 PM