Dlp Mcp

io.github.cx-anand-nandeshwar由 GitHub 账号 cx-anand-nandeshwar 发布v1.0.0更新于 Nov 2, 2026

The DLP MCP provides the compliance violation in the one drive, google drive documents.

已验证Streamable HTTP可网页运行开发者工具安全与监控数据与分析

概览

AI 生成的概览

将 AI 助手连接到 DLP One 应用安全平台,用于发起扫描、查看漏洞发现并获取修复建议。

可以做什么
  • 规划、触发并监控 SAST、SCA、IaC 与密钥检测扫描,按严重级别筛选列出扫描与发现项
  • 查看单条发现详情并更新风险结果状态
  • 创建、搜索和配置项目,并将项目归入应用
  • 读取租户级漏洞汇总、风险评分与按时间窗口的趋势
  • 为代码、存在漏洞的依赖包和容器镜像生成 AI 修复建议
适用场景
适合已在使用 DLP One 的团队,希望助手在 IDE 或 AI 代理内直接发起扫描、分诊漏洞并起草修复方案,而不必切换到平台网页界面。
运行要求
需要 DLP API 主机上带租户路径的远程 HTTP 端点。需要 DLP One 租户和有效的 DLP 凭据:在 Authorization 请求头中提供 API 密钥,或通过客户端完成 OAuth2 登录。需要能访问 DLP 主机的网络。
安装前请注意
该服务器拥有对代码和仓库的安全扫描权限:触发扫描可能将本地代码或仓库地址发送到 DLP One。它会创建项目和应用并更改发现项状态,写入会影响租户内的共享数据。Authorization 请求头中的 API 密钥属于需妥善保管的凭据;OAuth2 会将客户端跳转到 DLP One 登录页。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Dlp Mcp,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

其他 MCP 客户端

参照 仓库 中的启动说明。

README

DLP MCP Server

[License]

A production-ready Model Context Protocol (MCP) server that connects AI coding assistants to DLP One — enabling real-time security scanning, vulnerability management, and AI-generated remediation directly inside your IDE or AI agent.

Table of Contents

Overview

The DLP Security MCP server bridges your AI assistant (Claude, Cursor, Copilot, etc.) with DLP One's enterprise application security platform. It exposes security workflows as natural-language-accessible MCP tools, allowing developers to scan code, investigate findings, and receive context-aware fixes without leaving their development environment.

Supported scan engines:

  • SAST — Static Application Security Testing (30+ languages)
  • SCA — Software Composition Analysis (open-source dependencies)
  • KICS — Infrastructure as Code security (Terraform, CloudFormation, Kubernetes, Dockerfile)
  • Secret Detection — Hardcoded credentials, API keys, tokens

Supported transport protocols

Multi-protocol support to facilitate secure and efficient communication between clients and the server:

  • stdio: Standard input/output for CLI or embedded agents
  • sse (Server-Sent Events): For real-time streaming to web-based clients
  • httpstreamableHttp: HTTP-compatible protocol for stream-based messaging

Features

CategoryCapabilities
ScanningPlan, trigger, and monitor multi-engine security scans (CLI or API mode)
FindingsList, filter, and inspect vulnerabilities with severity and state tracking
RemediationAI-generated fixes for code vulnerabilities, insecure packages, and container images
Project ManagementCreate, configure, and search DLP One projects
Application ManagementGroup projects into applications and get org-wide security metrics
AnalyticsTenant-wide vulnerability summaries, risk scores, and time-windowed trends
Supply ChainDetect malicious npm/Maven/PyPI/Go/NuGet packages via Dustico integration
Enterprise AuthJWT (JWKS-verified), OAuth2 token exchange, Redis session caching
ObservabilityStructured logging (zerolog), OpenTelemetry tracing

Authentication

The server uses API Key and OAuth2 authentication.

API Key Authentication

  1. Clients authenticate to DLP One and get an API key.
  2. This API key will be used during MCP client configuration, include the API Key in the Authorization header as mentioned in the MCP Client Configuration section.

OAuth2 Authentication

DLP MCP supports Dynamic Client Registration (DCR) flow allows an AI client (such as Cursor or Claude Desktop) to connect securely.

  1. User only needs to configure the MCP client as mentioned in the MCP Client Configuration section.
  2. When the client attempts to connect to the MCP server, it will be redirected to DLP One login page for authentication.
  3. Once authentication is successful with valid DLP credentials, the MCP client can use the tools provided by the MCP server.

Note: You required valid DLP credentials to get the API Key or connect to the MCP server.

Refer Authentication for detailed authentication instructions and troubleshooting.

MCP Client Configuration

Prerequisites

  • A DLP One tenant
  • DLP API Host
  • API Key (with required access if using API key authentication)

JSON Configuration

Below are examples to add the server to your MCP client configuration. See the examples/ folder for ready-to-use client config files.

Cursor IDE

API Key Authentication:

json
{  "mcpServers": {    "DLP": {      "url": "https://{api_host}/api/security-mcp/mcp/{tenant}",      "headers": {        "cx-origin": "Curosr",        "Authorization": "API_KEY"      }    }  }}

OAuth2 Authentication

json
{  "mcpServers": {    "DLP": {      "url": "https://{api_host}/api/security-mcp/mcp/{tenant}"    }  }}
Claude Desktop / Claude Code

API Key Authentication:

json
{  "mcpServers": {    "DLP": {      "type": "http",      "url": "https://{api_host}/api/security-mcp/mcp/{tenant}",      "headers": {        "Authorization": "<API_KEY>"      }    }  }}

Available Tools

Refer usage for detail information.

Scanning

ToolDescription
planScanRecommend scan engines based on the project
triggerScanStart a scan (CLI for local code, API for repository URL)
getScanDetailsGet scan status, progress, and severity summary
getLatestScansRetrieve recent scans for a project
listScansList scans with status, date, and branch filters
listFindingsList vulnerabilities from a scan with severity filtering
getFindingDetailsGet detailed information for a specific finding

Project management

ToolDescription
resolveProjectLook up a project by name
createProjectCreate a new DLP One project
listProjectsBrowse or search all projects
getProjectConfigGet full project configuration

Application management

ToolDescription
listApplicationsBrowse or search applications
createApplicationCreate a new application
getApplicationDetailsGet application details by ID
associateProjectLink projects to an application

Analytics & risk

ToolDescription
listProjectsOverviewHigh-level project security status
getProjectsOverviewAggregateOrganization-wide project metrics
getApplicationsOverviewAggregateOrganization-wide application metrics
getTenantVulnerabilitiesSummaryTime-windowed vulnerability analytics
getRiskSummaryOverall risk assessment
listRiskResultsRisk findings with filtering
updateRiskResultStatusUpdate finding status (CONFIRMED, URGENT, NOT_EXPLOITABLE, …)

Remediation

ToolDescription
codeRemediationAI-generated fixes for SAST findings, secrets, and IaC misconfigurations
packageRemediationSafe upgrade paths for vulnerable open-source packages
imageRemediationSecure base image alternatives for vulnerable container images

License

Apache 2.0 — see LICENSE for details.

Contributing

See CONTRIBUTING.md for development setup, module architecture, and contribution guidelines.

Website: DLP.

© 2026 DLP Ltd. All Rights Reserved.

来源:README.md,提交 5e4696a

工具

0
SourceWeft 尚未读取此服务器的工具。

版本历史

1
  1. v1.0.0最新Sep 16, 2026

评分

已满足适用检查项的 50%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分

    请保持端点能够响应 MCP 握手;SourceWeft 每 30 天探测一次。

  • 已声明认证方式必需10 分

    请在注册中心条目中声明凭据请求头,或保持端点能够响应探测,以便判定其认证方式。

  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分