upstash-vector-js

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: References the official @upstash/vector SDK for interacting with Upstash Vector services. These are vendor-provided resources from a recognized technology service provider.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and retrieval of user-provided data through its vector operations, creating a surface where retrieved data could potentially be interpreted as instructions if processed by a downstream agent without sanitization.
  • Ingestion points: upsert, query, and fetch methods located in sdk-methods.md and features/filtering-and-metadata.md.
  • Boundary markers: The instructions do not provide specific prompt boundary markers for data retrieved from the database.
  • Capability inventory: Network access to the Upstash Vector API; execution of vector and metadata queries.
  • Sanitization: The skill relies on the Upstash API's internal handling of structured metadata and vector data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 12:38 AM
Security Audit — agent-trust-hub — upstash-vector-js