Agent Skill

io.github.Deep-Identity-Inc由 GitHub 账号 Deep-Identity-Inc 发布v1.0.1更新于 Oct 25, 2026

An MCP Server that provides identity verification and anti-fraud tools for AI agents via deepidv.

已验证Streamable HTTP可网页运行金融安全与监控

概览

AI 生成的概览

远程 MCP 服务器，通过 deepidv 提供身份验证与反欺诈操作，例如申请人、验证会话、筛查和银行流水数据

可以做什么
搜索申请人、查看资料、检查邀请状态并重新发送邀请。列出并查看验证会话，获取产物和时间线，创建或使会话过期，并更新结果。列出、查看并创建可复用的验证工作流。列出并获取银行流水请求与详情，查看统计并按外部 ID 查询。运行 PEP、制裁、头衔和负面媒体筛查，并轮询异步任务。
适用场景
当助手需要基于现有 deepidv 账户管理身份验证会话、申请人、合规筛查或银行流水请求，而不想单独开发集成时适用。
运行要求
位于 mcp.deepidv.com 的远程 streamable HTTP 端点，要求 OAuth。托管服务器使用共享的公共 OAuth 客户端，并以 deepidv 邮箱、密码及必要时的 MFA 验证用户。需要 deepidv 账户。
安装前请注意
该服务器处理敏感的身份与财务数据，包括身份证件、自拍、银行流水和筛查结果。多个工具会创建或修改数据：创建和使会话过期、更新结果、创建工作流、重新发送邀请以及创建银行流水请求。登录需要 deepidv 账户凭据和 MFA。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Agent Skill，将其添加到工作区。
  2. 连接 所需的 凭据。它们会被私密存储，绝不会在公开页面上展示。
  3. 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。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体，或 SourceWeft 对话中。智能体会读取该服务器的安装指南，先向你展示发布者、端点和登录方式，经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.Deep-Identity-Inc%2Fagent-skill/install.md，并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前，先告诉我它的发布者、端点和登录方式，等我确认。不要把 API 密钥写进命令或我会提交的文件；问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。 客户端会在首次连接时完成 OAuth 登录。

{
  "mcpServers": {
    "agent-skill": {
      "type": "http",
      "url": "https://mcp.deepidv.com/v1/mcp"
    }
  }
}

README

deepidv: AI-Powered Verification Engine

[License: Apache-2.0] [Skill Standard] [GDPR Compliant] [SOC 2 Compliant] [CCPA Compliant]

Verify anyone, anywhere in seconds. deepidv is an AI-powered verification engine that replaces the fragmented stack of identity vendors with a single platform. ID verification, face liveness, deepfake detection, AML screening, background checks, and more, all from one API.

This repository makes those capabilities accessible to AI agents and brings agentic intelligence into the deepidv platform itself. Three workstreams define how.

The Workstreams

Verify Skill

Any compatible AI agent can load this skill from the repository and immediately call deepidv's verification and screening APIs with no separate integration required. Endpoint routing, authentication, and request formatting are all handled by the skill.

  • Face liveness — Confirms a selfie came from a live person, not a photo or replay attack
  • Identity verification — Matches a government-issued ID against a selfie and validates the document
  • Deepfake detection — Determines whether an image or video is AI-generated or synthetically manipulated
  • Adverse media screening — Searches global news and public records for negative risk signals tied to a person
  • AML and sanctions screening — Checks a subject against PEP lists, sanctions databases, and financial crime watchlists
  • Combined verification — Runs a complete verification and screening flow in a single API call

Compatible with Claude Code, Codex, Cursor, Windsurf, and OpenCode. Full invocation guidance, endpoint routing, authentication instructions, and worked examples are in skills/deepidv-verify/SKILL.md.

MCP Server

The MCP server exposes hosted DeepIDV platform operations as structured tools that MCP-compatible clients can call directly. It is available at https://mcp.deepidv.com/v1/mcp and supports remote OAuth-based access.

  • Applicants - Search applicants, inspect profiles, check invitation status, and resend invites
  • Verification sessions - List sessions, inspect session details, retrieve artifacts, review timelines, create new verification sessions, expire sessions, and update outcomes
  • Workflows - List workflows, inspect workflow definitions, and create reusable workflows
  • Financial requests - List bank statements, retrieve bank statement details, view stats, query by external ID, and create bank statement requests
  • Silent screening - Run PEP and sanctions checks, title checks, adverse-media screening, and async job polling

Hosted MCP access uses a shared public OAuth client:

  • client_id - deepidv if the client explicitly asks for it
  • client_secret - not used
  • user authentication - deepidv email, password, and MFA when required

Setup and usage docs are in mcp-server/README.md and mcp-server/llms-install.md.

deepAI Assistant Skill

The deepAI Assistant is a coding agent skill that activates in context when you're building a deepidv integration. Rather than bouncing between docs and your editor, you get SDK guidance, error explanations, webhook setup help, and compliance workflow advice right in your editor.

The assistant loads automatically when a coding agent detects deepidv SDK usage, API calls, webhook configuration, or KYC workflow implementation in the active codebase. It is aligned to the published @deepidv/server TypeScript docs and the public API surfaces so it can steer developers toward documented integration patterns.

See skills/deepai-assistant/SKILL.md for the current integration workflow, auth guidance, and troubleshooting notes.

Marketplace Distribution

The Verify skill is listed across the major agent skill directories and MCP marketplaces, including SkillsMP, SkillHub, Agent-Skills.md, MCP Market, Skills Directory, LobeHub Skills, Orthogonal, and Awesome MCP Servers.

If you maintain a skill directory and would like to list this skill, open an issue or reach out to [email protected].

Contributing

Bug reports, corrections to reference documentation, and improvements to skill definitions are welcome. Open an issue to discuss before submitting a pull request for anything substantial.

About deepidv

deepidv is a full-stack verification engine built to replace the fragmented stack of identity providers most teams are stuck with. It verifies identities, runs compliance screening, and monitors for fraud continuously. One platform, one API, one dashboard, one team.

Learn more at deepidv.com.

License

Apache License 2.0

来源：README.md，提交 0dda96d

工具

0
该服务器需要登录，因此 SourceWeft 无法在没有账号的情况下读取其工具，也未收录。已连接账号所能看到的内容可能因账号及其权限而异。

版本历史

1
  1. v1.0.1最新Sep 16, 2026

评分

已满足适用检查项的 85%

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
  • 已声明认证方式必需10 分
  • 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