ReasoningBank with AgentDB

Pass

Audited by Gen Agent Trust Hub on Apr 2, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is composed of a single documentation file (SKILL.md) providing instructions for integrating with AgentDB. No executable code or malicious scripts are included in the skill distribution.
  • [EXTERNAL_DOWNLOADS]: The instructions recommend using npx agentdb@latest to initialize the database and set up the MCP server. This involves fetching the package from the standard npm registry.
  • [COMMAND_EXECUTION]: The documentation provides standard shell commands for database initialization, migration, and management. These commands are typical for the described functionality of setting up a local vector database.
  • [PROMPT_INJECTION]: The skill describes a system that ingests user queries and agent trajectories into a database for retrieval. This creates a surface for indirect prompt injection where data processed at runtime could influence agent behavior, but the static instructions do not contain any malicious injection patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 2, 2026, 07:18 AM
Security Audit — agent-trust-hub — ReasoningBank with AgentDB