Entropy0

Name: Entropy0
Author: entropy0dev

@entropy0dev

关于 Entropy0

Source trust for AI agents — evaluates URLs and domains before ingestion, returns proceed/deny decisions with scores and reason codes.

#mcp #trust #security #ai-agents #url-safety

基本信息

分类

其他

传输方式

stdio

发布者

entropy0dev

提交者

Entropy0

配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "entropy0": {
      "command": "npx",
      "args": [
        "-y",
        "@entropy0/mcp"
      ],
      "env": {
        "ENTROPY0_API_KEY": "<YOUR_API_KEY>"
      }
    }
  }
}

工具

未检测到工具

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is Entropy0?

Entropy0 is an MCP server that provides trust and URL safety tools for AI agents. It acts as a trust gate between the agent and the external web, returning a machine-readable recommended action before the agent fetches, navigates, or transacts. It's designed for developers building agents that interact with external URLs in environments like Claude Desktop, Cursor, and Cline.

How to use Entropy0?

Install the MCP server package via npm (@entropy0/mcp) and configure it with your Entropy0 API key. The server exposes tools that evaluate domains or URLs, returning a recommended action that your agent can reason about. Configuration details are managed through the MCP integration in your client application.

Key features of Entropy0

Deterministic pipeline using WHOIS, DNS, SSL, and threat intel feeds
Auditable, explainable, and overridable decisions with confidence clamps
Low-confidence negatives never hard-deny
Returns recommended_action, reason codes, and bounded validity window

Use cases of Entropy0

Trust-gate every URL before an agent fetches its content
Evaluate candidate sources in a LangGraph retrieval pipeline
Block requests to flagged domains via Express middleware
Prevent attribution of claims to sources where article body text was not captured

FAQ from Entropy0

How does Entropy0 make decisions?

Each decision runs a deterministic pipeline that classifies the domain (Clear Threat → Safe Known) using registered data, maps it to a base action under your chosen policy, shifts strictness based on interaction risk, and applies confidence clamps. Same inputs always produce the same output.