vector-database

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides legitimate technical documentation and implementation templates for vector database operations, including schema design, index tuning, and model selection.
  • [SAFE]: Code examples and scripts follow industry best practices for security, such as utilizing environment variables for database credentials and API keys.
  • [SAFE]: The skill includes a dedicated 'RAG Ingest Pipeline Checklist' that explicitly mandates PII redaction and multi-tenancy isolation for data ingestion, mitigating risks of data exposure.
  • [SAFE]: All external libraries and tools referenced (OpenAI, Weaviate, Cohere, pgvector) are well-known, established services for vector database and AI workflows.
  • [SAFE]: No evidence of prompt injection, obfuscation, or unauthorized data exfiltration was found across the skill's instructions and reference files.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 01:53 PM
Security Audit — agent-trust-hub — vector-database