agent-zone-mcp
MCP server for Agent Zone — vendor-neutral infrastructure knowledge, K8s validation, and execution templates for AI agents. 200+ articles, 10 tools, no API key required.
README
agent-zone-mcp
Infrastructure Knowledge MCP Server
MCP server for Agent Zone — vendor-neutral infrastructure knowledge, K8s validation, and execution templates for AI agents.
Search 200+ articles, validate K8s manifests, get structured playbooks, and discover the right validation path for your infrastructure work. Free, no API key required.
Features
- 200+ infrastructure articles — Kubernetes, networking, observability, CI/CD, disaster recovery, security, multi-region patterns
- 10 MCP tools — search, validate, get templates and playbooks, submit feedback
- K8s manifest validation — 5 built-in policy checks (resource limits, health probes, security, labels, image tags)
- 5 validation paths — from static analysis (~40% detection) to cloud ephemeral (~98%)
- 19 infrastructure templates — searchable by path, tags, and query with download support
- 6 structured playbooks — step-by-step operational guidance filtered by validation path
- Vendor-neutral — not tied to any cloud provider
- Free, no API key — works out of the box
Quick Start
npx agent-zone-mcp
Claude Code
claude mcp add agent-zone -- npx agent-zone-mcp
Claude Desktop / Cursor / Windsurf
Add to your MCP config:
{
"mcpServers": {
"agent-zone": {
"command": "npx",
"args": ["-y", "agent-zone-mcp"]
}
}
}
Streamable HTTP
For clients that support HTTP transport directly:
https://api.agent-zone.ai/mcp
Tools
Knowledge
| Tool | Description |
|---|---|
search |
Full-text search across all articles. Filter by category. |
get_article |
Retrieve full article content by ID (markdown + metadata). |
list_categories |
Browse all knowledge categories with article counts. |
submit_feedback |
Report helpful, inaccurate, or outdated content. |
suggest_topic |
Suggest missing topics for the knowledge base. |
Templates & Playbooks
| Tool | Description |
|---|---|
list_templates |
Search/filter infrastructure templates by validation path, tags, or query. |
get_template |
Get full template metadata including requirements and produces. |
get_playbook |
Get a structured step-by-step playbook, optionally filtered by validation path. |
Validation
| Tool | Description |
|---|---|
select_validation_path |
Describe your resources and needs, get the right validation path + templates + playbook. |
validate |
Submit K8s manifests or Helm values, get policy validation results back. |
How It Works
This is a stdio-based MCP server that wraps the Agent Zone REST API. The npm package runs locally and communicates with api.agent-zone.ai over HTTPS. No API key, no account, no local database — just install and use.
A Streamable HTTP endpoint is also available at https://api.agent-zone.ai/mcp for clients that support remote MCP servers directly.
Example: End-to-End Validation Flow
An agent needs to test a Helm chart and has Docker with 8GB RAM:
-
select_validation_path({work_type: "helm-chart", available_resources: ["docker"], resource_specs: {memory_mb: 8192}})→ recommends Path 3 (minikube) with matching templates and thevalidate-helm-chartplaybook -
list_templates({path: 3, query: "minikube"})→ finds setup templates -
get_template({id: "local-full~minikube~profiles"})→ gets requirements and setup details -
get_playbook({id: "validate-helm-chart", path: 3})→ structured steps with commands and success criteria -
validate({type: "k8s-manifest", content: "apiVersion: apps/v1\nkind: Deployment..."})→ instant policy checks (resource limits, health probes, security, labels)
Validation Paths
| Path | Name | Requirements | Detection Rate |
|---|---|---|---|
| 1 | Static Analysis | None | ~40% |
| 2 | Local Lightweight | Docker | ~75% |
| 3 | Local Full-Fidelity | Docker + 4GB+ RAM | ~90% |
| 4 | Cloud Ephemeral | Cloud account | ~98% |
| 5 | Free-Tier Cloud | Codespaces/free tier | ~85% |
Policy Checks
The validate tool checks K8s manifests against 5 policies:
- require-resource-limits — CPU and memory limits on all containers
- require-health-checks — liveness and readiness probes
- no-latest-tag — pinned image tags (no
:latestor missing tag) - no-privileged — no privileged mode, root user, or privilege escalation
- require-labels — standard
app.kubernetes.io/nameandapp.kubernetes.io/versionlabels
Links
License
MIT
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。