LLM Tools Guide

Reviews, benchmarks, and security assessments for AI agent tooling

Agent Card Documentation MCP Tools Security Training

Quick Start

# Connect via MCP
{
  "mcpServers": {
    "llm-tools": {
      "url": "https://llm-tools-guide.dev/mcp",
      "transport": "streamable-http"
    }
  }
}

# Or discover via A2A protocol
curl https://llm-tools-guide.dev/.well-known/agent.json

Tool Categories

CategoryTop PickSecurity RatingReview
Code GenerationClaude Code91/100Review
Agent FrameworksLangChain67/100Review
MCP ServersFilesystem87/100Review
Web BrowsingPuppeteer MCP54/100Review
MemoryMemory MCP58/100Review

Security Assessment

Every tool reviewed here is tested against AgentPwn's 48 attack scenarios. Scores reflect real-world resilience against prompt injection, data exfiltration, and capability abuse.

API Endpoints

MethodPathDescription
GET/.well-known/agent.jsonA2A agent card
POST/a2a/tasks/sendSubmit a task
POST/mcp/tools/listList available MCP tools
POST/mcp/tools/callExecute an MCP tool
GET/contextBrowse documentation