# python-sdk - Doramagic AI Context Pack

> 定位：安装前体验与判断资产。它帮助宿主 AI 有一个好的开始，但不代表已经安装、执行或验证目标项目。

## 充分原则

- **充分原则，不是压缩原则**：AI Context Pack 应该充分到让宿主 AI 在开工前理解项目价值、能力边界、使用入口、风险和证据来源；它可以分层组织，但不以最短摘要为目标。
- **压缩策略**：只压缩噪声和重复内容，不压缩会影响判断和开工质量的上下文。

## 给宿主 AI 的使用方式

你正在读取 Doramagic 为 python-sdk 编译的 AI Context Pack。请把它当作开工前上下文：帮助用户理解适合谁、能做什么、如何开始、哪些必须安装后验证、风险在哪里。不要声称你已经安装、运行或执行了目标项目。

## Claim 消费规则

- **事实来源**：Repo Evidence + Claim/Evidence Graph；Human Wiki 只提供显著性、术语和叙事结构。
- **事实最低状态**：`supported`
- `supported`：可以作为项目事实使用，但回答中必须引用 claim_id 和证据路径。
- `weak`：只能作为低置信度线索，必须要求用户继续核实。
- `inferred`：只能用于风险提示或待确认问题，不能包装成项目事实。
- `unverified`：不得作为事实使用，应明确说证据不足。
- `contradicted`：必须展示冲突来源，不得替用户强行选择一个版本。

## 它最适合谁

- **正在使用 Claude/Codex/Cursor/Gemini 等宿主 AI 的开发者**：README 或插件配置提到多个宿主 AI。 证据：`README.md` Claim：`clm_0002` supported 0.86

## 它能做什么

- **命令行启动或安装流程**（需要安装后验证）：项目文档中存在可执行命令，真实使用需要在本地或宿主环境中运行这些命令。 证据：`README.md` Claim：`clm_0001` supported 0.86

## 怎么开始

- `pip install "mcp[cli]"` 证据：`README.md` Claim：`clm_0003` supported 0.86
- `claude mcp add --transport http my-server http://localhost:8000/mcp` 证据：`README.md` Claim：`clm_0004` supported 0.86
- `npx -y @modelcontextprotocol/inspector` 证据：`README.md` Claim：`clm_0005` supported 0.86

## 继续前判断卡

- **当前建议**：需要管理员/安全审批
- **为什么**：继续前可能涉及密钥、账号、外部服务或敏感上下文，建议先经过管理员或安全审批。

### 30 秒判断

- **现在怎么做**：需要管理员/安全审批
- **最小安全下一步**：先跑 Prompt Preview；若涉及凭证或企业环境，先审批再试装
- **先别相信**：角色质量和任务匹配不能直接相信。
- **继续会触碰**：角色选择偏差、命令执行、宿主 AI 配置

### 现在可以相信

- **适合人群线索：正在使用 Claude/Codex/Cursor/Gemini 等宿主 AI 的开发者**（supported）：有 supported claim 或项目证据支撑，但仍不等于真实安装效果。 证据：`README.md` Claim：`clm_0002` supported 0.86
- **能力存在：命令行启动或安装流程**（supported）：可以相信项目包含这类能力线索；是否适合你的具体任务仍要试用或安装后验证。 证据：`README.md` Claim：`clm_0001` supported 0.86
- **存在 Quick Start / 安装命令线索**（supported）：可以相信项目文档出现过启动或安装入口；不要因此直接在主力环境运行。 证据：`README.md` Claim：`clm_0003` supported 0.86

### 现在还不能相信

- **角色质量和任务匹配不能直接相信。**（unverified）：角色库证明有很多角色，不证明每个角色都适合你的具体任务，也不证明角色能产生高质量结果。
- **不能把角色文案当成真实执行能力。**（unverified）：安装前只能判断角色描述和任务画像是否匹配，不能证明它能在宿主 AI 里完成任务。
- **真实输出质量不能在安装前相信。**（unverified）：Prompt Preview 只能展示引导方式，不能证明真实项目中的结果质量。
- **宿主 AI 版本兼容性不能在安装前相信。**（unverified）：Claude、Cursor、Codex、Gemini 等宿主加载规则和版本差异必须在真实环境验证。
- **不会污染现有宿主 AI 行为，不能直接相信。**（inferred）：Skill、plugin、AGENTS/CLAUDE/GEMINI 指令可能改变宿主 AI 的默认行为。 证据：`AGENTS.md`, `CLAUDE.md`
- **可安全回滚不能默认相信。**（unverified）：除非项目明确提供卸载和恢复说明，否则必须先在隔离环境验证。
- **真实安装后是否与用户当前宿主 AI 版本兼容？**（unverified）：兼容性只能通过实际宿主环境验证。
- **项目输出质量是否满足用户具体任务？**（unverified）：安装前预览只能展示流程和边界，不能替代真实评测。

### 继续会触碰什么

- **角色选择偏差**：用户对任务应该由哪个专家角色处理的判断。 原因：选错角色会让 AI 从错误专业视角回答，浪费时间或误导决策。
- **命令执行**：包管理器、网络下载、本地插件目录、项目配置或用户主目录。 原因：运行第一条命令就可能产生环境改动；必须先判断是否值得跑。 证据：`README.md`
- **宿主 AI 配置**：Claude/Codex/Cursor/Gemini/OpenCode 等宿主的 plugin、Skill 或规则加载配置。 原因：宿主配置会改变 AI 后续工作方式，可能和用户已有规则冲突。 证据：`AGENTS.md`, `CLAUDE.md`
- **本地环境或项目文件**：安装结果、插件缓存、项目配置或本地依赖目录。 原因：安装前无法证明写入范围和回滚方式，需要隔离验证。 证据：`README.md`
- **环境变量 / API Key**：项目入口文档明确出现 API key、token、secret 或账号凭证配置。 原因：如果真实安装需要凭证，应先使用测试凭证并经过权限/合规判断。 证据：`README.md`, `README.v2.md`, `examples/clients/simple-chatbot/README.MD`
- **宿主 AI 上下文**：AI Context Pack、Prompt Preview、Skill 路由、风险规则和项目事实。 原因：导入上下文会影响宿主 AI 后续判断，必须避免把未验证项包装成事实。

### 最小安全下一步

- **先跑 Prompt Preview**：先用交互式试用验证任务画像和角色匹配，不要先导入整套角色库。（适用：任何项目都适用，尤其是输出质量未知时。）
- **只在隔离目录或测试账号试装**：避免安装命令污染主力宿主 AI、真实项目或用户主目录。（适用：存在命令执行、插件配置或本地写入线索时。）
- **先备份宿主 AI 配置**：Skill、plugin、规则文件可能改变 Claude/Cursor/Codex 的默认行为。（适用：存在插件 manifest、Skill 或宿主规则入口时。）
- **不要使用真实生产凭证**：环境变量/API key 一旦进入宿主或工具链，可能产生账号和合规风险。（适用：出现 API、TOKEN、KEY、SECRET 等环境线索时。）
- **安装后只验证一个最小任务**：先验证加载、兼容、输出质量和回滚，再决定是否深用。（适用：准备从试用进入真实工作流时。）

### 退出方式

- **保留安装前状态**：记录原始宿主配置和项目状态，后续才能判断是否可恢复。
- **准备移除宿主 plugin / Skill / 规则入口**：如果试装后行为异常，可以把宿主 AI 恢复到试装前状态。
- **保留原始角色选择记录**：如果输出偏题，可以回到任务画像阶段重新选择角色，而不是继续沿着错误角色推进。
- **记录安装命令和写入路径**：没有明确卸载说明时，至少要知道哪些目录或配置需要手动清理。
- **准备撤销测试 API key 或 token**：测试凭证泄露或误用时，可以快速止损。
- **如果没有回滚路径，不进入主力环境**：不可回滚是继续前阻断项，不应靠信任或运气继续。

## 哪些只能预览

- 解释项目适合谁和能做什么
- 基于项目文档演示典型对话流程
- 帮助用户判断是否值得安装或继续研究

## 哪些必须安装后验证

- 真实安装 Skill、插件或 CLI
- 执行脚本、修改本地文件或访问外部服务
- 验证真实输出质量、性能和兼容性

## 边界与风险判断卡

- **把安装前预览误认为真实运行**：用户可能高估项目已经完成的配置、权限和兼容性验证。 处理方式：明确区分 prompt_preview_can_do 与 runtime_required。 Claim：`clm_0006` inferred 0.45
- **命令执行会修改本地环境**：安装命令可能写入用户主目录、宿主插件目录或项目配置。 处理方式：先在隔离环境或测试账号中运行。 证据：`README.md` Claim：`clm_0007` supported 0.86
- **待确认**：真实安装后是否与用户当前宿主 AI 版本兼容？。原因：兼容性只能通过实际宿主环境验证。
- **待确认**：项目输出质量是否满足用户具体任务？。原因：安装前预览只能展示流程和边界，不能替代真实评测。
- **待确认**：安装命令是否需要网络、权限或全局写入？。原因：这影响企业环境和个人环境的安装风险。

## 开工前工作上下文

### 加载顺序

- 先读取 how_to_use.host_ai_instruction，建立安装前判断资产的边界。
- 读取 claim_graph_summary，确认事实来自 Claim/Evidence Graph，而不是 Human Wiki 叙事。
- 再读取 intended_users、capabilities 和 quick_start_candidates，判断用户是否匹配。
- 需要执行具体任务时，优先查 role_skill_index，再查 evidence_index。
- 遇到真实安装、文件修改、网络访问、性能或兼容性问题时，转入 risk_card 和 boundaries.runtime_required。

### 任务路由

- **命令行启动或安装流程**：先说明这是安装后验证能力，再给出安装前检查清单。 边界：必须真实安装或运行后验证。 证据：`README.md` Claim：`clm_0001` supported 0.86

### 上下文规模

- 文件总数：437
- 重要文件覆盖：40/437
- 证据索引条目：48
- 角色 / Skill 条目：37

### 证据不足时的处理

- **missing_evidence**：说明证据不足，要求用户提供目标文件、README 段落或安装后验证记录；不要补全事实。
- **out_of_scope_request**：说明该任务超出当前 AI Context Pack 证据范围，并建议用户先查看 Human Manual 或真实安装后验证。
- **runtime_request**：给出安装前检查清单和命令来源，但不要替用户执行命令或声称已执行。
- **source_conflict**：同时展示冲突来源，标记为待核实，不要强行选择一个版本。

## Prompt Recipes

### 适配判断

- 目标：判断这个项目是否适合用户当前任务。
- 预期输出：适配结论、关键理由、证据引用、安装前可预览内容、必须安装后验证内容、下一步建议。

```text
请基于 python-sdk 的 AI Context Pack，先问我 3 个必要问题，然后判断它是否适合我的任务。回答必须包含：适合谁、能做什么、不能做什么、是否值得安装、证据来自哪里。所有项目事实必须引用 evidence_refs、source_paths 或 claim_id。
```

### 安装前体验

- 目标：让用户在安装前感受核心工作流，同时避免把预览包装成真实能力或营销承诺。
- 预期输出：一段带边界标签的体验剧本、安装后验证清单和谨慎建议；不含真实运行承诺或强营销表述。

```text
请把 python-sdk 当作安装前体验资产，而不是已安装工具或真实运行环境。

请严格输出四段：
1. 先问我 3 个必要问题。
2. 给出一段“体验剧本”：用 [安装前可预览]、[必须安装后验证]、[证据不足] 三种标签展示它可能如何引导工作流。
3. 给出安装后验证清单：列出哪些能力只有真实安装、真实宿主加载、真实项目运行后才能确认。
4. 给出谨慎建议：只能说“值得继续研究/试装”“先补充信息后再判断”或“不建议继续”，不得替项目背书。

硬性边界：
- 不要声称已经安装、运行、执行测试、修改文件或产生真实结果。
- 不要写“自动适配”“确保通过”“完美适配”“强烈建议安装”等承诺性表达。
- 如果描述安装后的工作方式，必须使用“如果安装成功且宿主正确加载 Skill，它可能会……”这种条件句。
- 体验剧本只能写成“示例台词/假设流程”：使用“可能会询问/可能会建议/可能会展示”，不要写“已写入、已生成、已通过、正在运行、正在生成”。
- Prompt Preview 不负责给安装命令；如用户准备试装，只能提示先阅读 Quick Start 和 Risk Card，并在隔离环境验证。
- 所有项目事实必须来自 supported claim、evidence_refs 或 source_paths；inferred/unverified 只能作风险或待确认项。

```

### 角色 / Skill 选择

- 目标：从项目里的角色或 Skill 中挑选最匹配的资产。
- 预期输出：候选角色或 Skill 列表，每项包含适用场景、证据路径、风险边界和是否需要安装后验证。

```text
请读取 role_skill_index，根据我的目标任务推荐 3-5 个最相关的角色或 Skill。每个推荐都要说明适用场景、可能输出、风险边界和 evidence_refs。
```

### 风险预检

- 目标：安装或引入前识别环境、权限、规则冲突和质量风险。
- 预期输出：环境、权限、依赖、许可、宿主冲突、质量风险和未知项的检查清单。

```text
请基于 risk_card、boundaries 和 quick_start_candidates，给我一份安装前风险预检清单。不要替我执行命令，只说明我应该检查什么、为什么检查、失败会有什么影响。
```

### 宿主 AI 开工指令

- 目标：把项目上下文转成一次对话开始前的宿主 AI 指令。
- 预期输出：一段边界明确、证据引用明确、适合复制给宿主 AI 的开工前指令。

```text
请基于 python-sdk 的 AI Context Pack，生成一段我可以粘贴给宿主 AI 的开工前指令。这段指令必须遵守 not_runtime=true，不能声称项目已经安装、运行或产生真实结果。
```


## 角色 / Skill 索引

- 共索引 37 个角色 / Skill / 项目文档条目。

- **Development Guidelines**（project_doc）：- main is currently the V2 rework. Breaking changes are expected here — when removing or replacing an API, delete it outright and document the change in docs/migration.md . Do not add @deprecated shims or backward-compat layers on main . - v1.x is the release branch for the current stable line. Backport PRs target this branch and use a v1.x title prefix. - README.md is frozen at v1 a pre-commit hook rejects edits .… 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`AGENTS.md`
- **Claude**（project_doc）： 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CLAUDE.md`
- **MCP Python SDK**（project_doc）：Python implementation of the Model Context Protocol MCP 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`README.md`
- **Python SDK Examples**（project_doc）：This folders aims to provide simple examples of using the Python SDK. Please refer to the servers repository https://github.com/modelcontextprotocol/servers for real-world servers. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/README.md`
- **Simple Auth Client Example**（project_doc）：A demonstration of how to use the MCP Python SDK with OAuth authentication over streamable HTTP or SSE transport. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/clients/simple-auth-client/README.md`
- **MCP Simple Chatbot**（project_doc）：This example demonstrates how to integrate the Model Context Protocol MCP into a simple CLI chatbot. The implementation showcases MCP's flexibility by supporting multiple tools through MCP servers and is compatible with any LLM provider that follows OpenAI API standards. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/clients/simple-chatbot/README.MD`
- **Simple Task Client**（project_doc）：A minimal MCP client demonstrating polling for task results over streamable HTTP. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/clients/simple-task-client/README.md`
- **Simple Interactive Task Client**（project_doc）：A minimal MCP client demonstrating responses to interactive tasks elicitation and sampling . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/clients/simple-task-interactive-client/README.md`
- **MCP SSE Polling Demo Client**（project_doc）：Demonstrates client-side auto-reconnect for the SSE polling pattern SEP-1699 . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/clients/sse-polling-client/README.md`
- **MCP Everything Server**（project_doc）：A comprehensive MCP server implementing all protocol features for conformance testing. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/everything-server/README.md`
- **MCP OAuth Authentication Demo**（project_doc）：This example demonstrates OAuth 2.0 authentication with the Model Context Protocol using separate Authorization Server AS and Resource Server RS to comply with the new RFC 9728 specification. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-auth/README.md`
- **MCP Simple Pagination**（project_doc）：A simple MCP server demonstrating pagination for tools, resources, and prompts using cursor-based pagination. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-pagination/README.md`
- **MCP Simple Prompt**（project_doc）：A simple MCP server that exposes a customizable prompt template with optional context and topic parameters. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-prompt/README.md`
- **MCP Simple Resource**（project_doc）：A simple MCP server that exposes sample text files as resources. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-resource/README.md`
- **MCP Simple StreamableHttp Stateless Server Example**（project_doc）：MCP Simple StreamableHttp Stateless Server Example 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-streamablehttp-stateless/README.md`
- **MCP Simple StreamableHttp Server Example**（project_doc）：MCP Simple StreamableHttp Server Example 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-streamablehttp/README.md`
- **Simple Interactive Task Server**（project_doc）：A minimal MCP server demonstrating interactive tasks with elicitation and sampling. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-task-interactive/README.md`
- **Simple Task Server**（project_doc）：A minimal MCP server demonstrating the experimental tasks feature over streamable HTTP. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-task/README.md`
- **Usage**（project_doc）：A simple MCP server that exposes a website fetching tool. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/simple-tool/README.md`
- **MCP SSE Polling Demo Server**（project_doc）：Demonstrates the SSE polling pattern with server-initiated stream close for long-running tasks SEP-1699 . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`examples/servers/sse-polling-demo/README.md`
- **Contributing**（project_doc）：Thank you for your interest in contributing to the MCP Python SDK! This document provides guidelines and instructions for contributing. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CONTRIBUTING.md`
- **Authorization**（project_doc）：This page is currently being written. Check back soon for complete documentation. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/authorization.md`
- **Concepts**（project_doc）：This page is currently being written. Check back soon for complete documentation. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/concepts.md`
- **MCP Python SDK**（project_doc）：The Model Context Protocol MCP allows applications to provide context for LLMs in a standardized way, separating the concerns of providing context from the actual LLM interaction. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/index.md`
- **Installation**（project_doc）：The Python SDK is available on PyPI as mcp https://pypi.org/project/mcp/ so installation is as simple as: 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/installation.md`
- **Low-Level Server**（project_doc）：This page is currently being written. Check back soon for complete documentation. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/low-level-server.md`
- **Migration Guide: v1 to v2**（project_doc）：This guide covers the breaking changes introduced in v2 of the MCP Python SDK and how to update your code. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/migration.md`
- **Experimental Features**（project_doc）：The features in this section are experimental and may change without notice. They track the evolving MCP specification and are not yet stable. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/experimental/index.md`
- **Client Task Usage**（project_doc）：Tasks are an experimental feature. The API may change without notice. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/experimental/tasks-client.md`
- **Server Task Implementation**（project_doc）：Tasks are an experimental feature. The API may change without notice. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/experimental/tasks-server.md`
- **Tasks**（project_doc）：Tasks are an experimental feature tracking the draft MCP specification. The API may change without notice. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/experimental/tasks.md`
- **Testing MCP Servers**（project_doc）：The Python SDK provides a Client class for testing MCP servers with an in-memory transport. This makes it easy to write tests without network overhead. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/testing.md`
- **Contributor Covenant Code of Conduct**（project_doc）：Contributor Covenant Code of Conduct 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CODE_OF_CONDUCT.md`
- **MCP Python SDK**（project_doc）：Python implementation of the Model Context Protocol MCP 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`README.v2.md`
- **Release Process**（project_doc）：1. Change dependency version in pyproject.toml 2. Upgrade lock with uv lock --resolution lowest-direct 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`RELEASE.md`
- **Security Policy**（project_doc）：Thank you for helping keep the Model Context Protocol and its ecosystem secure. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`SECURITY.md`
- **Step 1: Resolve the PR**（project_doc）：Review the pull request: $ARGUMENTS 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`.claude/commands/review-pr.md`

## 证据索引

- 共索引 48 条证据。

- **Development Guidelines**（documentation）：- main is currently the V2 rework. Breaking changes are expected here — when removing or replacing an API, delete it outright and document the change in docs/migration.md . Do not add @deprecated shims or backward-compat layers on main . - v1.x is the release branch for the current stable line. Backport PRs target this branch and use a v1.x title prefix. - README.md is frozen at v1 a pre-commit hook rejects edits . Edit README.v2.md instead. 证据：`AGENTS.md`
- **Claude**（documentation）：@AGENTS.md 证据：`CLAUDE.md`
- **MCP Python SDK**（documentation）：Python implementation of the Model Context Protocol MCP 证据：`README.md`
- **Python SDK Examples**（documentation）：This folders aims to provide simple examples of using the Python SDK. Please refer to the servers repository https://github.com/modelcontextprotocol/servers for real-world servers. 证据：`examples/README.md`
- **Simple Auth Client Example**（documentation）：A demonstration of how to use the MCP Python SDK with OAuth authentication over streamable HTTP or SSE transport. 证据：`examples/clients/simple-auth-client/README.md`
- **MCP Simple Chatbot**（documentation）：This example demonstrates how to integrate the Model Context Protocol MCP into a simple CLI chatbot. The implementation showcases MCP's flexibility by supporting multiple tools through MCP servers and is compatible with any LLM provider that follows OpenAI API standards. 证据：`examples/clients/simple-chatbot/README.MD`
- **Simple Task Client**（documentation）：A minimal MCP client demonstrating polling for task results over streamable HTTP. 证据：`examples/clients/simple-task-client/README.md`
- **Simple Interactive Task Client**（documentation）：A minimal MCP client demonstrating responses to interactive tasks elicitation and sampling . 证据：`examples/clients/simple-task-interactive-client/README.md`
- **MCP SSE Polling Demo Client**（documentation）：Demonstrates client-side auto-reconnect for the SSE polling pattern SEP-1699 . 证据：`examples/clients/sse-polling-client/README.md`
- **MCP Everything Server**（documentation）：A comprehensive MCP server implementing all protocol features for conformance testing. 证据：`examples/servers/everything-server/README.md`
- **MCP OAuth Authentication Demo**（documentation）：This example demonstrates OAuth 2.0 authentication with the Model Context Protocol using separate Authorization Server AS and Resource Server RS to comply with the new RFC 9728 specification. 证据：`examples/servers/simple-auth/README.md`
- **MCP Simple Pagination**（documentation）：A simple MCP server demonstrating pagination for tools, resources, and prompts using cursor-based pagination. 证据：`examples/servers/simple-pagination/README.md`
- **MCP Simple Prompt**（documentation）：A simple MCP server that exposes a customizable prompt template with optional context and topic parameters. 证据：`examples/servers/simple-prompt/README.md`
- **MCP Simple Resource**（documentation）：A simple MCP server that exposes sample text files as resources. 证据：`examples/servers/simple-resource/README.md`
- **MCP Simple StreamableHttp Stateless Server Example**（documentation）：MCP Simple StreamableHttp Stateless Server Example 证据：`examples/servers/simple-streamablehttp-stateless/README.md`
- **MCP Simple StreamableHttp Server Example**（documentation）：MCP Simple StreamableHttp Server Example 证据：`examples/servers/simple-streamablehttp/README.md`
- **Simple Interactive Task Server**（documentation）：A minimal MCP server demonstrating interactive tasks with elicitation and sampling. 证据：`examples/servers/simple-task-interactive/README.md`
- **Simple Task Server**（documentation）：A minimal MCP server demonstrating the experimental tasks feature over streamable HTTP. 证据：`examples/servers/simple-task/README.md`
- **Usage**（documentation）：A simple MCP server that exposes a website fetching tool. 证据：`examples/servers/simple-tool/README.md`
- **MCP SSE Polling Demo Server**（documentation）：Demonstrates the SSE polling pattern with server-initiated stream close for long-running tasks SEP-1699 . 证据：`examples/servers/sse-polling-demo/README.md`
- **Contributing**（documentation）：Thank you for your interest in contributing to the MCP Python SDK! This document provides guidelines and instructions for contributing. 证据：`CONTRIBUTING.md`
- **License**（source_file）：Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files the "Software" , to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: 证据：`LICENSE`
- **Authorization**（documentation）：This page is currently being written. Check back soon for complete documentation. 证据：`docs/authorization.md`
- **Concepts**（documentation）：This page is currently being written. Check back soon for complete documentation. 证据：`docs/concepts.md`
- **MCP Python SDK**（documentation）：The Model Context Protocol MCP allows applications to provide context for LLMs in a standardized way, separating the concerns of providing context from the actual LLM interaction. 证据：`docs/index.md`
- **Installation**（documentation）：The Python SDK is available on PyPI as mcp https://pypi.org/project/mcp/ so installation is as simple as: 证据：`docs/installation.md`
- **Low-Level Server**（documentation）：This page is currently being written. Check back soon for complete documentation. 证据：`docs/low-level-server.md`
- **Migration Guide: v1 to v2**（documentation）：This guide covers the breaking changes introduced in v2 of the MCP Python SDK and how to update your code. 证据：`docs/migration.md`
- **Experimental Features**（documentation）：The features in this section are experimental and may change without notice. They track the evolving MCP specification and are not yet stable. 证据：`docs/experimental/index.md`
- **Client Task Usage**（documentation）：Tasks are an experimental feature. The API may change without notice. 证据：`docs/experimental/tasks-client.md`
- **Server Task Implementation**（documentation）：Tasks are an experimental feature. The API may change without notice. 证据：`docs/experimental/tasks-server.md`
- **Tasks**（documentation）：Tasks are an experimental feature tracking the draft MCP specification. The API may change without notice. 证据：`docs/experimental/tasks.md`
- **Testing MCP Servers**（documentation）：The Python SDK provides a Client class for testing MCP servers with an in-memory transport. This makes it easy to write tests without network overhead. 证据：`docs/testing.md`
- **Contributor Covenant Code of Conduct**（documentation）：Contributor Covenant Code of Conduct 证据：`CODE_OF_CONDUCT.md`
- **MCP Python SDK**（documentation）：Python implementation of the Model Context Protocol MCP 证据：`README.v2.md`
- **Release Process**（documentation）：1. Change dependency version in pyproject.toml 2. Upgrade lock with uv lock --resolution lowest-direct 证据：`RELEASE.md`
- **Security Policy**（documentation）：Thank you for helping keep the Model Context Protocol and its ecosystem secure. 证据：`SECURITY.md`
- **Step 1: Resolve the PR**（documentation）：Review the pull request: $ARGUMENTS 证据：`.claude/commands/review-pr.md`
- **Servers Config**（structured_config）：{ "mcpServers": { "sqlite": { "command": "uvx", "args": "mcp-server-sqlite", "--db-path", "./test.db" }, "puppeteer": { "command": "npx", "args": "-y", "@modelcontextprotocol/server-puppeteer" } } } 证据：`examples/clients/simple-chatbot/mcp_simple_chatbot/servers_config.json`
- **Applied 120 line-length rule to all files: https://github.com/modelcontextprotocol/python-sdk/pull/856**（source_file）：Applied 120 line-length rule to all files: https://github.com/modelcontextprotocol/python-sdk/pull/856 543961968c0634e93d919d509cce23a1d6a56c21 证据：`.git-blame-ignore-revs`
- **Generated**（source_file）：Generated uv.lock linguist-generated=true 证据：`.gitattribute`
- **Dependabot**（source_file）：version: 2 updates: - package-ecosystem: "github-actions" directory: "/" schedule: interval: monthly groups: github-actions: patterns: - " " 证据：`.github/dependabot.yml`
- **Byte-compiled / optimized / DLL files**（source_file）：Byte-compiled / optimized / DLL files pycache / .py cod $py.class 证据：`.gitignore`
- **TODO Max : Drop this in v2.**（source_file）：repos: - repo: https://github.com/pre-commit/pre-commit-hooks rev: v6.0.0 hooks: - id: end-of-file-fixer 证据：`.pre-commit-config.yaml`
- **TODO Marcelo : Add Anthropic copyright?**（source_file）：site name: MCP Server site description: MCP Server strict: true 证据：`mkdocs.yml`
- **PEP 517 build isolation fetches build-system .requires and transitives at**（source_file）：project name = "mcp" dynamic = "version" description = "Model Context Protocol SDK" readme = "README.md" requires-python = " =3.10" authors = { name = "Model Context Protocol a Series of LF Projects, LLC." } maintainers = { name = "David Soria Parra", email = "davidsp@anthropic.com" }, { name = "Marcelo Trylesinski", email = "marcelotryle@gmail.com" }, { name = "Max Isbey", email = "maxisbey@anthropic.com" }, { name = "Felix Weinberger", email = "fweinberger@anthropic.com" }, keywords = "git", "mcp", "llm", "automation" license = { text = "MIT" } classifiers = "Development Status :: 4 - Beta", "Intended Audience :: Developers", "License :: OSI Approved :: MIT License", "Programming Language… 证据：`pyproject.toml`
- **!/usr/bin/env python3**（source_file）：!/usr/bin/env python3 """Update README.md with live code snippets from example files. 证据：`scripts/update_readme_snippets.py`
- **Conftest**（source_file）：@pytest.fixture def anyio backend : return "asyncio" 证据：`tests/conftest.py`

## 宿主 AI 必须遵守的规则

- **把本资产当作开工前上下文，而不是运行环境。**：AI Context Pack 只包含证据化项目理解，不包含目标项目的可执行状态。 证据：`AGENTS.md`, `CLAUDE.md`, `README.md`
- **回答用户时区分可预览内容与必须安装后才能验证的内容。**：安装前体验的消费者价值来自降低误装和误判，而不是伪装成真实运行。 证据：`AGENTS.md`, `CLAUDE.md`, `README.md`

## 用户开工前应该回答的问题

- 你准备在哪个宿主 AI 或本地环境中使用它？
- 你只是想先体验工作流，还是准备真实安装？
- 你最在意的是安装成本、输出质量、还是和现有规则的冲突？

## 验收标准

- 所有能力声明都能回指到 evidence_refs 中的文件路径。
- AI_CONTEXT_PACK.md 没有把预览包装成真实运行。
- 用户能在 3 分钟内看懂适合谁、能做什么、如何开始和风险边界。

---

## Doramagic Context Augmentation

The following material strengthens the Repomix/AI Context Pack body. Human Manual is only a reading skeleton; pitfall logs become hard operating constraints for the host AI.

## Human Manual Skeleton

Usage rule: this is only a reading path and salience signal, not factual authority. Concrete facts must still come from repo evidence / Claim Graph.

Hard rules for the host AI:
- Do not treat page titles, order, summaries, or importance as project facts.
- When explaining the Human Manual skeleton, state that it is only a reading path / salience signal.
- Capability, installation, compatibility, runtime status, and risk judgments must cite repo evidence, source paths, or Claim Graph.

- **Getting Started with MCP Python SDK**：importance `high`
  - source_paths: pyproject.toml, README.md, examples/snippets/servers/fastmcp_quickstart.py
- **Project Structure**：importance `high`
  - source_paths: src/mcp/__init__.py, src/mcp/client/__init__.py, src/mcp/server/__init__.py, src/mcp/shared/__init__.py, src/mcp/types/__init__.py
- **FastMCP Server Development**：importance `high`
  - source_paths: src/mcp/server/fastmcp/__init__.py, src/mcp/server/fastmcp/server.py, src/mcp/server/fastmcp/tools/__init__.py, src/mcp/server/fastmcp/resources/__init__.py, src/mcp/server/fastmcp/prompts/__init__.py
- **Low-Level Server Development**：importance `medium`
  - source_paths: src/mcp/server/lowlevel/server.py, src/mcp/server/lowlevel/__init__.py, src/mcp/server/session.py, examples/snippets/servers/lowlevel/basic.py, examples/snippets/servers/lowlevel/lifespan.py
- **Server Lifecycle and Context Management**：importance `high`
  - source_paths: src/mcp/server/fastmcp/context.py, src/mcp/server/context.py, src/mcp/server/mcpserver/context.py, examples/snippets/servers/lifespan_example.py
- **Resources**：importance `high`
  - source_paths: src/mcp/server/fastmcp/resources/base.py, src/mcp/server/fastmcp/resources/resource_manager.py, src/mcp/server/fastmcp/resources/templates.py, examples/snippets/servers/basic_resource.py
- **Tools and Structured Output**：importance `high`
  - source_paths: src/mcp/server/fastmcp/tools/base.py, src/mcp/server/fastmcp/tools/tool_manager.py, src/mcp/server/validation.py, examples/snippets/servers/structured_output.py, examples/snippets/servers/direct_call_tool_result.py
- **Prompts and Templates**：importance `medium`
  - source_paths: src/mcp/server/fastmcp/prompts/base.py, src/mcp/server/fastmcp/prompts/manager.py, examples/snippets/servers/basic_prompt.py, examples/servers/simple-prompt/mcp_simple_prompt/server.py

## Repo Inspection Evidence

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `e8e64842781c66b613872cf394de6e7d6f6925bf`
- inspected_files: `pyproject.toml`, `README.md`, `uv.lock`, `docs/authorization.md`, `docs/concepts.md`, `docs/index.md`, `docs/low-level-server.md`, `docs/installation.md`, `docs/testing.md`, `docs/migration.md`, `docs/hooks/gen_ref_pages.py`, `docs/experimental/tasks.md`, `docs/experimental/index.md`, `docs/experimental/tasks-client.md`, `docs/experimental/tasks-server.md`, `examples/README.md`, `examples/snippets/pyproject.toml`, `examples/mcpserver/readme-quickstart.py`, `examples/mcpserver/simple_echo.py`, `examples/mcpserver/logging_and_progress.py`

Hard rules for the host AI:
- Without repo_clone_verified=true, do not claim the source code has been read.
- Without repo_inspection_verified=true, do not turn README/docs/package observations into facts.
- Without quick_start_verified=true, do not claim the Quick Start has been successfully run.

## Doramagic Pitfall Constraints

These rules come from Doramagic discovery, validation, or compilation pitfalls. The host AI must treat them as operating constraints, not general background notes.

### Constraint 1: 来源证据：Duplicate `initialize` with changed parameters can overwrite `ServerSession.client_params`

- Trigger: GitHub 社区证据显示该项目存在一个配置相关的待验证问题：Duplicate `initialize` with changed parameters can overwrite `ServerSession.client_params`
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | cevd_fbd28a768e3d4de5b6350068d5f755e6 | https://github.com/modelcontextprotocol/python-sdk/issues/2605 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 2: 来源证据：Streamable HTTP server silently drops in-flight request when client reuses a JSON-RPC id

- Trigger: GitHub 社区证据显示该项目存在一个配置相关的待验证问题：Streamable HTTP server silently drops in-flight request when client reuses a JSON-RPC id
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | cevd_4ac932bd22b544cdb3fd360948cea40a | https://github.com/modelcontextprotocol/python-sdk/issues/2655 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 3: 来源证据：streamable_http_client: one concurrent request HTTPStatusError tears down sibling requests

- Trigger: GitHub 社区证据显示该项目存在一个配置相关的待验证问题：streamable_http_client: one concurrent request HTTPStatusError tears down sibling requests
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | cevd_eea700d7503c4f4aa8ee4dcbd3db4591 | https://github.com/modelcontextprotocol/python-sdk/issues/2604 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 4: 来源证据：FastMCP crashes when tool return type uses Python 3.10+ `A | B | C` union syntax

- Trigger: GitHub 社区证据显示该项目存在一个维护/版本相关的待验证问题：FastMCP crashes when tool return type uses Python 3.10+ `A | B | C` union syntax
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能阻塞安装或首次运行。
- Evidence: community_evidence:github | cevd_a8726c44efdb4ae880d844a7d492b1e9 | https://github.com/modelcontextprotocol/python-sdk/issues/2591 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 5: 来源证据：Feature Proposal: Secure Tool/Resource/Prompt Decorators with Auth + Encrypted I/O

- Trigger: GitHub 社区证据显示该项目存在一个安全/权限相关的待验证问题：Feature Proposal: Secure Tool/Resource/Prompt Decorators with Auth + Encrypted I/O
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能影响授权、密钥配置或安全边界。
- Evidence: community_evidence:github | cevd_8f0ce3f22c2e45b9af5ad37292daba8b | https://github.com/modelcontextprotocol/python-sdk/issues/1305 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 6: 来源证据：Progress notifications cause server to hang on stdio transport

- Trigger: GitHub 社区证据显示该项目存在一个安全/权限相关的待验证问题：Progress notifications cause server to hang on stdio transport
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能阻塞安装或首次运行。
- Evidence: community_evidence:github | cevd_ed924e0ca71b419381464d8c25d237ef | https://github.com/modelcontextprotocol/python-sdk/issues/1141 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 7: 来源证据：Streamable HTTP server accepts mismatched `MCP-Protocol-Version` header and body `protocolVersion` on `initialize`

- Trigger: GitHub 社区证据显示该项目存在一个安全/权限相关的待验证问题：Streamable HTTP server accepts mismatched `MCP-Protocol-Version` header and body `protocolVersion` on `initialize`
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能影响授权、密钥配置或安全边界。
- Evidence: community_evidence:github | cevd_84eac7ce86a345c38a09308517763962 | https://github.com/modelcontextprotocol/python-sdk/issues/2618 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 8: 来源证据：User-Agent header in sHTTP transport is not forwarded to auth flow

- Trigger: GitHub 社区证据显示该项目存在一个安全/权限相关的待验证问题：User-Agent header in sHTTP transport is not forwarded to auth flow
- Host AI rule: 来源问题仍为 open，Pack Agent 需要复核是否仍影响当前版本。
- Why it matters: 可能影响授权、密钥配置或安全边界。
- Evidence: community_evidence:github | cevd_824f98c734544a28b3bfd8e982425150 | https://github.com/modelcontextprotocol/python-sdk/issues/1664 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 9: 失败模式：installation: Bug: `anyio.Lock` in `oauth2.py` raises "current task is not holding this lock" under cross-t...

- Trigger: Developers should check this installation risk before relying on the project: Bug: `anyio.Lock` in `oauth2.py` raises "current task is not holding this lock" under cross-task generator driving
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Bug: `anyio.Lock` in `oauth2.py` raises "current task is not holding this lock" under cross-task generator driving. Context: Observed when using windows
- Why it matters: Developers may fail before the first successful local run: Bug: `anyio.Lock` in `oauth2.py` raises "current task is not holding this lock" under cross-task generator driving
- Evidence: failure_mode_cluster:github_issue | fmev_7ab763a43233ec1e7dcaebc3255017a6 | https://github.com/modelcontextprotocol/python-sdk/issues/2644 | Bug: `anyio.Lock` in `oauth2.py` raises "current task is not holding this lock" under cross-task generator driving
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.

### Constraint 10: 失败模式：installation: FastMCP.__init__ clobbers root logger with INFO RichHandler — hangs stdio servers under back-...

- Trigger: Developers should check this installation risk before relying on the project: FastMCP.__init__ clobbers root logger with INFO RichHandler — hangs stdio servers under back-pressure
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: FastMCP.__init__ clobbers root logger with INFO RichHandler — hangs stdio servers under back-pressure. Context: Observed when using node, python, macos
- Why it matters: Developers may fail before the first successful local run: FastMCP.__init__ clobbers root logger with INFO RichHandler — hangs stdio servers under back-pressure
- Evidence: failure_mode_cluster:github_issue | fmev_4529ab733a6558a19e9c3f318ce112c5 | https://github.com/modelcontextprotocol/python-sdk/issues/2527 | FastMCP.__init__ clobbers root logger with INFO RichHandler — hangs stdio servers under back-pressure
- Hard boundary: do not present this pitfall as solved, verified, or safe to ignore unless later validation evidence explicitly closes it.