Trust OS MCP Server
MCP server for verifying high-impact decisions with Trust OS.
README
Trust OS MCP Server
MCP server for verifying high-impact decisions with Trust OS.
Connect any MCP-compatible AI agent — Claude Desktop, Cursor, or any MCP client — to the Trust OS Decision Verification API and verify financial and operational decisions before execution.
Links
- Website: https://trust-os.io
- Docs: https://trust-os.io/docs
- API Reference: https://trust-os.io/docs/api
- Playground: https://demo.trust-os.io
- Operations Demo: https://ops.trust-os.io
- OpenAPI: https://trust-os.io/openapi.json
- GitHub Org: https://github.com/trustos-trustfolio
What it does
The Trust OS MCP Server exposes the Trust OS Decision Verification API as an MCP tool. When an AI agent calls verify_decision, the server:
- Validates the input (action, amount, currency, destination, etc.)
- POSTs the payload to Trust OS
/v1/decision/verify - Returns a structured response including risk score, recommendation, and cryptographic proof
Installation
git clone https://github.com/trustos-trustfolio/trustos-mcp-server.git
cd trustos-mcp-server
npm install
npm run build
Environment Variables
Copy .env.example to .env and fill in your API key:
cp .env.example .env
| Variable | Required | Description |
|---|---|---|
TRUSTOS_API_KEY |
Yes | Your Trust OS API key |
TRUSTOS_BASE_URL |
No | Override API base URL (default: Trust OS gateway) |
Never commit .env to version control.
Claude Desktop Configuration
Add the following to your Claude Desktop claude_desktop_config.json:
{
"mcpServers": {
"trustos": {
"command": "node",
"args": ["C:/trustos-mcp-server/dist/index.js"],
"env": {
"TRUSTOS_API_KEY": "YOUR_API_KEY"
}
}
}
}
See examples/claude_desktop_config.json for the full example.
Claude Desktop config file location:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Tool: verify_decision
Description: Verify a high-impact decision before execution using Trust OS.
Input schema:
| Field | Type | Required | Description |
|---|---|---|---|
action |
string | Yes | The action or decision to verify |
amount |
number | No | Monetary amount involved |
currency |
string | No | Currency code (e.g. USDC, USD) |
destination |
string | No | Destination address or identifier |
source |
string | No | Source address or identifier |
priority |
string | No | Priority level (e.g. high, critical) |
metadata |
object | No | Additional context as key-value pairs |
Example Prompt
Use Trust OS to verify this decision before execution:
action stablecoin_transfer, amount 50000, currency USDC, destination wallet_abc.
Example Response
Trust OS Decision Verification Result
════════════════════════════════════════
decision_id : txn_a1b2c3d4
recommendation: APPROVE
risk_score : 12
risk_level : LOW
policy : standard_transfer_policy
proof_hash : 0xabc123...
verified : true
latency_ms : 142
Full response:
{
"decision_id": "txn_a1b2c3d4",
"recommendation": "APPROVE",
"risk_score": 12,
...
}
Security
- Store
TRUSTOS_API_KEYin the MCP client'senvconfig, never in source code .envis git-ignored- API key is never logged or included in error messages
- To report a vulnerability, email founder@trust-os.io (see SECURITY.md)
Contributing
See CONTRIBUTING.md.
Changelog
See CHANGELOG.md.
License
MIT © Trustfolio Inc.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。