pgvector

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely instructional and contains no executable scripts, shell commands, or remote code downloads. It provides standard SQL templates for database configuration and retrieval.- [DATA_EXPOSURE]: The skill explicitly promotes security best practices by instructing users to filter by tenant_id and user_id within SQL queries to ensure data isolation in multi-tenant environments.- [INDIRECT_PROMPT_INJECTION]: While the skill describes how to store and retrieve unstructured text (agent memories), it recommends using parameterized queries and structured schemas, reducing the risk of injection attacks during database operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 01:03 PM
Security Audit — agent-trust-hub — pgvector