开源 AI 治理工具。每个工具既可作为独立的 MCP 服务器运行，也可作为 CLI 使用 — 它们共同构成了一个治理网格。
由 Maiife 构建 — 企业级 AI 控制平面。
软件包 | 描述 | 已发布 |
| 所有工具包使用的共享类型和格式化程序 | ✅ |
| AI 环境扫描器 — 发现 IDE 扩展、MCP 服务器、智能体框架、API 密钥、本地模型 | ✅ |
| MCP 服务器安全扫描器 — 对权限、数据敏感度、影响范围进行配置评分 | ✅ |
| “你的 AI 技术栈是什么？” — 可分享的 AI 工具包个人资料卡 | ✅ |
| MCP 健康检查与自动修复工具 — 你的 MCP 设置的“全科医生” | ✅ |
| 个人 AI 使用日记 — 追踪你的 AI 使用情况，获取反思性见解 | ✅ |
| 跨工具 AI 记忆同步 — 一个 context.json，同步至 Cursor、Claude、MCP | ✅ |
| 提示词质量分析器 — 对你的 AI 提示词进行评分、改进和 lint 检查 | ✅ |
| LLM-as-judge 评估引擎 — 使用结构化准则对智能体输出进行评分 | ✅ |
| 智能体工作流追踪器 — 追踪、查看和分析执行跨度 | ✅ |
| AI 支出计算器 + 优化器 — 跨供应商的统一成本报告 | ✅ |
| 游戏化提示词教练 — 通过等级、连胜、徽章来提升提示词水平 | ✅ |
| 个人 AI 订阅审计员 — 发现 AI 支出中的浪费 | ✅ |
| 个人模型推荐器 — 为你的任务找到最合适的模型 | ✅ |
| AI 周度回顾 — 每周为你生成一份 AI 使用情况的“Spotify Wrapped”报告 | ✅ |
Related MCP server: Mund
# Scan your AI environment
npx @maiife-ai-pub/probe scan
# Audit your MCP server security
npx @maiife-ai-pub/mcp-audit scan
# Generate your AI Stack profile card
npx @maiife-ai-pub/ai-stack --format svg --output my-stack.svg
# Health check your MCP servers
npx @maiife-ai-pub/mcp-doctor check
# Log an AI interaction
npx @maiife-ai-pub/ai-journal log --tool claude --task coding --duration 30
# Sync AI context across tools
npx @maiife-ai-pub/context-sync push
# Score your AI prompts
npx @maiife-ai-pub/prompt-score analyze --input prompt.txt
# Evaluate agent outputs with rubrics
npx @maiife-ai-pub/eval score --rubric code-review --input review.txt
# Trace agent workflows
npx @maiife-ai-pub/trace list --days 7
# Track AI spend across vendors
npx @maiife-ai-pub/cost report --period last-30d
# Gamified prompt coaching
npx @maiife-ai-pub/prompt-craft score --input prompt.txt
# Audit AI subscriptions for waste
npx @maiife-ai-pub/sub-audit
# Find the best model for your tasks
npx @maiife-ai-pub/model-match recommend --task coding
# Generate your AI week in review
npx @maiife-ai-pub/weekly-ai-report generate每个带有 MCP 服务器的工具都可以添加到 Claude Desktop、Cursor 或任何兼容 MCP 的客户端中。每个工具都通过 stdio 传输公开。
~/Library/Application Support/Claude/claude_desktop_config.json (macOS) 或 %APPDATA%\Claude\claude_desktop_config.json (Windows)：
{
"mcpServers": {
"maiife-probe": {
"command": "npx",
"args": ["@maiife-ai-pub/probe", "mcp"]
},
"maiife-mcp-audit": {
"command": "npx",
"args": ["@maiife-ai-pub/mcp-audit", "mcp"]
},
"maiife-mcp-doctor": {
"command": "npx",
"args": ["@maiife-ai-pub/mcp-doctor", "mcp"]
},
"maiife-eval": {
"command": "npx",
"args": ["@maiife-ai-pub/eval", "mcp"]
},
"maiife-prompt-score": {
"command": "npx",
"args": ["@maiife-ai-pub/prompt-score", "mcp"]
},
"maiife-prompt-craft": {
"command": "npx",
"args": ["@maiife-ai-pub/prompt-craft", "mcp"]
},
"maiife-cost": {
"command": "npx",
"args": ["@maiife-ai-pub/cost", "mcp"]
},
"maiife-model-match": {
"command": "npx",
"args": ["@maiife-ai-pub/model-match", "mcp"]
},
"maiife-ai-stack": {
"command": "npx",
"args": ["@maiife-ai-pub/ai-stack", "mcp"]
},
"maiife-context-sync": {
"command": "npx",
"args": ["@maiife-ai-pub/context-sync", "mcp"]
},
"maiife-sub-audit": {
"command": "npx",
"args": ["@maiife-ai-pub/sub-audit", "mcp"]
},
"maiife-trace": {
"command": "npx",
"args": ["@maiife-ai-pub/trace", "mcp"]
}
}
}选择你需要的工具 — 无需全部添加。配置完成后，Claude 即可直接在聊天中调用 probe_scan、mcp_audit_scan、eval_score、prompt_score_analyze、cost_report 等工具。
每个 MCP 服务器都作为 Docker 镜像发布在 GHCR 上。适用于沙盒环境或 Glama 集成。
# Pull and run any server
docker run -i ghcr.io/sakthivelchan89/maiife-probe
docker run -i ghcr.io/sakthivelchan89/maiife-mcp-audit
docker run -i ghcr.io/sakthivelchan89/maiife-eval
# ... same pattern for all 12 packages
# Or build from source
docker build -f packages/probe/Dockerfile -t maiife-probe .
docker run -i maiife-probeDocker 镜像使用 stdio 传输（不暴露端口）。传递 -i 以便与 MCP 客户端进行交互式 stdin/stdout 通信。
本工具包旨在达到 MCP Tier 1 质量标准（根据 MCP SEP-1730）。其含义如下：
维度 | 状态 |
许可证 | Apache 2.0 — 规范的 SPDX，OSI 认证 |
传输 | 仅 stdio（无网络暴露） |
CI/CD | GitHub Actions：在 Node 18, 20, 22 上进行 lint + 类型检查 + 测试 |
测试覆盖率 | vitest + |
MCP 一致性 | 针对所有 12 个 MCP 服务器的协议合规性套件 |
安全扫描 | CodeQL（每周 + PR 时），Dependabot（每周） |
漏洞响应 | 48 小时内处理 CRITICAL / 7 天内处理 HIGH（见 DEPENDENCY_POLICY.md） |
问题分类 SLA | 2 个工作日（见 CONTRIBUTING.md） |
版本控制 | SemVer，跨包同步，CHANGELOG.md |
供应链 | 已提交 |
容器安全 | 非 root 用户，无暴露端口，GHCR 签名 |
本仓库中的每个 MCP 服务器都根据 MCP 协议契约进行了验证：
✅ stdio 传输不变性（stdout 上无非 JSON 输出）
✅ initialize 握手响应有效的 serverInfo + 功能列表
✅ tools/list 返回已记录的工具集
✅ 所有工具的 inputSchema 字段均为有效的 JSON Schema 对象
✅ 未知工具调用返回结构化错误（而非崩溃）
运行测试套件：
pnpm test:conformance # all packages
cd packages/probe && pnpm test:conformance # single packageSECURITY.md — 漏洞报告政策
CONTRIBUTING.md — 问题/PR 指南和 SLA
CHANGELOG.md — 版本历史（Keep a Changelog 格式）
DEPENDENCY_POLICY.md — 依赖选择标准与补丁 SLA
欢迎贡献！阅读 CONTRIBUTING.md 获取完整指南。简要版本：
Fork 本仓库
创建分支：git checkout -b feat/my-improvement
进行更改 — 每个包位于 packages/<name>/
运行测试：pnpm test && pnpm test:conformance
向 main 提交 PR — 描述你更改的内容及原因
请遵循现有的代码风格（TypeScript, ESM, Vitest 测试）。每个包在适用时应同时作为 CLI 和 MCP 服务器工作。
Apache 2.0 — 可免费使用、修改和分发。
Maiife 平台的一部分 — 企业级 AI 控制平面。
Scan the current environment for AI tools, MCP servers, agent frameworks, API keys, and local models
| Name | Required | Description | Default |
|---|---|---|---|
| path | No | Root path to scan (defaults to current directory) | |
| scope | No | Scan scope: full=everything, quick=IDE+MCP only, category=specific | full |
| categories | No | Comma-separated categories: ide,mcp,agents,keys,models,deps | |
| includeProjectDeps | No | Scan package.json/requirements.txt for AI dependencies |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must disclose behavioral traits. The word 'Scan' suggests a read operation, but there is no mention of side effects, permissions, safety, or potential impact on the environment. For a tool that scans files and possibly accesses sensitive data (API keys), this is a significant omission.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single sentence that front-loads the action and key details. Every word contributes to understanding the tool's purpose, with no filler or redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With no output schema and 4 optional parameters, the description fails to cover what the tool returns (e.g., a list of found items, JSON output). The agent lacks information on how to interpret results, which is critical for a scanning tool. Additionally, it does not explain the behavior of different scopes or categories beyond what the schema provides.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the baseline is 3. The description does not add any additional meaning beyond the schema; it simply restates the categories listed in the 'categories' parameter description. No deeper semantics are provided.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action ('Scan') and the specific resources ('AI tools, MCP servers, agent frameworks, API keys, local models'), making it easy for an AI agent to understand the tool's purpose. No sibling tools exist, so differentiation is not required.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
While there are no sibling tools to compare against, the description implies the tool is for enumeration and discovery, which is sufficient. However, it lacks explicit guidance on when to use it (e.g., initial reconnaissance vs. targeted search), leaving some ambiguity.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
probe_scanprobe_scanv0.1.3probe_scanScored across 1 tool
Only one tool exists, so there is no possibility of ambiguity.
With a single tool, naming consistency is not applicable; the name 'probe_scan' is clear and descriptive.
One tool is too few for a toolkit; it feels thin and does not provide a meaningful set of capabilities.
The single scan tool likely misses complementary operations like listing previous scans, filtering, or exporting results, leaving the surface incomplete.
Security scanner for MCP servers. Detect vulnerabilities, prompt injection, and tool poisoning.
AI supply-chain BOM: agents, MCP servers, packages and recorded reach. Scan and investigate via MCP.
Scan any MCP server for tool-poisoning, security, auth & license. Trust score before install.
Find, compare, and audit software for AI agents. Scored registry of tools and MCP servers.