AgentDB Vector Search

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill documents several CLI commands using npx agentdb@latest. These are standard methods for interacting with the AgentDB CLI to initialize, query, and manage vector databases. These commands are intended for developers to run manually or for an agent to use as part of a development workflow.
  • [EXTERNAL_DOWNLOADS]: The skill references the agentdb and agentic-flow packages. Based on the author context (i-onlabs / ruvnet), these are official vendor resources for implementing vector search and agentic workflows. The downloads via npx target the official registry for these tools.
  • [REMOTE_CODE_EXECUTION]: The skill mentions npx agentdb@latest mcp to start a Model Context Protocol server. This is a standard integration pattern for modern AI agents (like Claude Code) to access specialized tools. The code executed is part of the official AgentDB package.
  • [DATA_EXFILTRATION]: While the skill involves managing and exporting vector data (e.g., agentdb export), these operations are local to the user's filesystem (./vectors.db, ./backup.json) and are standard database management tasks. No unauthorized external transmission of data was detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 05:03 AM
Security Audit — agent-trust-hub — AgentDB Vector Search