# mcp-grafana - 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 mcp-grafana 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_0004` supported 0.86
- **希望把专业流程带进宿主 AI 的用户**：仓库包含 Skill 文档。 证据：`.claude/skills/draft-release/SKILL.md` Claim：`clm_0005` supported 0.86

## 它能做什么

- **AI Skill / Agent 指令资产库**（可做安装前预览）：项目包含可被宿主 AI 读取的 Skill 或 Agent 指令文件，可用于把专业流程带入 Claude、Codex、Cursor 等宿主。 证据：`.claude/skills/draft-release/SKILL.md` Claim：`clm_0001` supported 0.86
- **多宿主安装与分发**（需要安装后验证）：项目包含插件或 marketplace 配置，说明它面向一个或多个 AI 宿主的安装和分发。 证据：`.claude-plugin/plugin.json` Claim：`clm_0002` supported 0.86
- **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_0003` supported 0.86

## 怎么开始

- `curl http://localhost:8000/healthz` 证据：`README.md` Claim：`clm_0006` supported 0.86
- `curl http://localhost:9090/healthz` 证据：`README.md` Claim：`clm_0007` 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**：工具权限边界不能在安装前相信。
- **继续会触碰**：命令执行、宿主 AI 配置、本地环境或项目文件

### 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_0004` supported 0.86
- **适合人群线索：希望把专业流程带进宿主 AI 的用户**（supported）：有 supported claim 或项目证据支撑，但仍不等于真实安装效果。 证据：`.claude/skills/draft-release/SKILL.md` Claim：`clm_0005` supported 0.86
- **能力存在：AI Skill / Agent 指令资产库**（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：`.claude/skills/draft-release/SKILL.md` Claim：`clm_0001` supported 0.86
- **能力存在：多宿主安装与分发**（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：`.claude-plugin/plugin.json` 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_0003` 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_0006` supported 0.86

### 现在还不能相信

- **工具权限边界不能在安装前相信。**（unverified）：MCP/tool 类项目通常会触碰文件、网络、浏览器或外部 API，必须真实检查权限和日志。
- **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：`.claude-plugin/plugin.json`, `.claude/skills/draft-release/SKILL.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。 证据：`.claude-plugin/plugin.json`
- **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。
- **安装命令是否需要网络、权限或全局写入？**（unverified）：这影响企业环境和个人环境的安装风险。 证据：`README.md`

### 继续会触碰什么

- **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：`.claude-plugin/plugin.json`, `.claude/skills/draft-release/SKILL.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：`.claude-plugin/plugin.json`, `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：`DEVELOPING.md`, `README.md`, `docs/docs.agent/context.md`, `docs/sources/_index.md` 等
- **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**：用安装前交互式试用判断工作方式是否匹配，不需要授权或改环境。（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。
- **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_0008` inferred 0.45
- **宿主 AI 插件或 Skill 规则冲突**：新规则可能改变用户现有宿主 AI 的工作方式。 处理方式：安装前先检查插件 manifest 和 Skill 文件，必要时隔离测试。 证据：`.claude-plugin/plugin.json` Claim：`clm_0009` supported 0.86
- **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_0010` 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

- **AI Skill / Agent 指令资产库**：先基于 role_skill_index / evidence_index 帮用户挑选可用角色、Skill 或工作流。 边界：可做安装前 Prompt 体验。 证据：`.claude/skills/draft-release/SKILL.md` Claim：`clm_0001` supported 0.86
- **多宿主安装与分发**：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：`.claude-plugin/plugin.json` Claim：`clm_0002` supported 0.86
- **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_0003` supported 0.86

### 上下文规模

- 文件总数：259
- 重要文件覆盖：40/259
- 证据索引条目：80
- 角色 / Skill 条目：1

### 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 mcp-grafana 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 mcp-grafana 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 mcp-grafana 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 1 个角色 / Skill / 项目文档条目。

- **draft-release**（skill）：Draft a new release. Bumps version, generates CHANGELOG.md from git history, and creates a PR. 激活提示：当用户任务与“draft-release”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.claude/skills/draft-release/SKILL.md`

## 证据索引

- 共索引 80 条证据。

- **docs.agent**（documentation）：An extremely minimal documentation AI agent to scope work, update context and plans, and author content. 证据：`docs/docs.agent/README.md`
- **Grafana MCP server**（documentation）：! Unit Tests https://github.com/grafana/mcp-grafana/actions/workflows/unit.yml/badge.svg https://github.com/grafana/mcp-grafana/actions/workflows/unit.yml ! Integration Tests https://github.com/grafana/mcp-grafana/actions/workflows/integration.yml/badge.svg https://github.com/grafana/mcp-grafana/actions/workflows/integration.yml ! E2E Tests https://github.com/grafana/mcp-grafana/actions/workflows/e2e.yml/badge.svg https://github.com/grafana/mcp-grafana/actions/workflows/e2e.yml ! Go Reference https://pkg.go.dev/badge/github.com/grafana/mcp-grafana.svg https://pkg.go.dev/github.com/grafana/mcp-grafana ! MCP Catalog https://archestra.ai/mcp-catalog/api/badge/quality/grafana/mcp-grafana https:… 证据：`README.md`
- **JSONSchema Linter**（documentation）：This linter helps detect and prevent a common issue with Go struct tags in this project. 证据：`internal/linter/jsonschema/README.md`
- **Troubleshooting**（documentation）：Common issues and solutions for mcp-grafana. 证据：`docs/troubleshooting.md`
- **Authentication**（documentation）：The Grafana MCP server needs credentials to call the Grafana API. Use a service account token recommended or a username and password. 证据：`docs/sources/configure/authentication.md`
- **Command-line flags**（documentation）：The mcp-grafana binary accepts flags for transports, tools, TLS, and observability. Run mcp-grafana --help for the exact list in your installed build. 证据：`docs/sources/configure/command-line-flags.md`
- **Multi-organization and headers**（documentation）：You can point the server at a specific Grafana organization and add custom HTTP headers to every request to Grafana. 证据：`docs/sources/configure/multi-organization-and-headers.md`
- **Server TLS streamable-http**（documentation）：When you use the streamable-http transport, you can serve the MCP server over HTTPS using your own TLS certificate and key. 证据：`docs/sources/configure/server-tls-streamable-http.md`
- **Transports and addresses**（documentation）：The Grafana MCP server supports three transports: stdio default for local use , SSE, and streamable-http. Choose the one that matches how your MCP client connects. 证据：`docs/sources/configure/transports-and-addresses.md`
- **Client configuration examples**（documentation）：This page walks through credentials, installation options, and MCP client JSON patterns for common editors and runtimes. 证据：`docs/sources/set-up/client-configuration-examples.md`
- **Deploy with Helm**（documentation）：Deploy the Grafana MCP server on Kubernetes using the Helm chart from the Grafana helm-charts repository. 证据：`docs/sources/set-up/deploy-with-helm.md`
- **Install with Docker**（documentation）：Run the Grafana MCP server using the official image from Docker Hub. The image defaults to SSE , but most users will want to use STDIO mode for direct integration with AI assistants like Claude Desktop. 证据：`docs/sources/set-up/install-with-docker.md`
- **Install with uvx**（documentation）：uv https://docs.astral.sh/uv/ is Astral's Python package manager and toolchain. uvx https://docs.astral.sh/uv/guides/tools/ runs a command from a published package in an isolated environment, without installing that package globally. It is similar to npx for Node.js. 证据：`docs/sources/set-up/install-with-uvx.md`
- **Grafana version compatibility**（documentation）：Some datasource API paths exist only in newer Grafana releases. This article explains a common error on Grafana before 9.0. 证据：`docs/sources/troubleshooting/grafana-version-compatibility.md`
- **Draft Release**（skill_instruction）：You are creating a new release for the mcp-grafana project. The user has provided a bump level: major , minor , or patch . 证据：`.claude/skills/draft-release/SKILL.md`
- **Plugin**（structured_config）：{ "name": "grafana", "version": "0.7.6", "description": "A Model Context Protocol MCP server for Grafana providing access to dashboards, datasources, and querying capabilities", "author": { "name": "Grafana Labs" }, "homepage": "https://github.com/grafana/mcp-grafana", "repository": "https://github.com/grafana/mcp-grafana", "license": "Apache-2.0", "mcpServers": { "grafana": { "command": "node", "args": "${CLAUDE PLUGIN ROOT}/.claude-plugin/install-binary.mjs" } } } 证据：`.claude-plugin/plugin.json`
- **License**（source_file）：Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ 证据：`LICENSE`
- **context.md**（documentation）：Follow and keep this documentation context and plan up to date. 证据：`docs/docs.agent/context.md`
- **docs agent**（documentation）：Assume the role of a technical writing and documentation agent for Grafana Labs. 证据：`docs/docs.agent/rules.md`
- **Grafana MCP server**（documentation）：This documentation helps you install the Grafana MCP server https://github.com/grafana/mcp-grafana , connect MCP-compatible clients, and configure authentication, transports, and tools. 证据：`docs/sources/_index.md`
- **Set up by client**（documentation）：Configure the Grafana MCP server for your MCP-compatible client. Each guide lists the config file location and a working server block. 证据：`docs/sources/clients/_index.md`
- **Claude Code**（documentation）：This guide helps you set up the mcp-grafana server for the Claude Code CLI. 证据：`docs/sources/clients/claude-code.md`
- **Claude Desktop**（documentation）：This guide helps you set up the mcp-grafana server for Claude Desktop. 证据：`docs/sources/clients/claude-desktop.md`
- **Codex CLI**（documentation）：This guide helps you set up the mcp-grafana server for the OpenAI Codex CLI. 证据：`docs/sources/clients/codex.md`
- **Cursor**（documentation）：This guide helps you set up the mcp-grafana server for Cursor. 证据：`docs/sources/clients/cursor.md`
- **Gemini CLI**（documentation）：This guide helps you set up the mcp-grafana server for the Google Gemini CLI. 证据：`docs/sources/clients/gemini-cli.md`
- **VS Code and GitHub Copilot**（documentation）：This guide helps you set up the mcp-grafana server for VS Code with GitHub Copilot agent mode. 证据：`docs/sources/clients/vscode-copilot.md`
- **Windsurf**（documentation）：This guide helps you set up the mcp-grafana server for Windsurf. 证据：`docs/sources/clients/windsurf.md`
- **Zed**（documentation）：This guide helps you set up the mcp-grafana server for the Zed editor. 证据：`docs/sources/clients/zed.md`
- **Configure the Grafana MCP server**（documentation）：Configure how the server connects to Grafana, which tools are enabled, and how clients connect to the server. 证据：`docs/sources/configure/_index.md`
- **Client TLS Grafana connection**（documentation）：If your Grafana instance uses mTLS or a custom CA, configure the MCP server to use the correct certificates when it calls the Grafana API. 证据：`docs/sources/configure/client-tls-grafana-connection.md`
- **Enable and disable tools**（documentation）：You can limit which tools the server exposes to reduce context window use or lock down capabilities and run the server in read-only mode. 证据：`docs/sources/configure/enable-and-disable-tools.md`
- **Health check endpoint**（documentation）：When you use the SSE -t sse or streamable HTTP -t streamable-http transport, the MCP server exposes a health check at /healthz . Load balancers, monitoring, and orchestration can use it to verify that the server is running and accepting connections. 证据：`docs/sources/configure/health-check-endpoint.md`
- **Proxied tools**（documentation）：Proxied tools are additional MCP tools that this server does not implement itself. It loads them from an MCP server that sits behind a Grafana datasource, using Grafana’s datasource proxy. Your client still talks only to this MCP server; the extra tools show up alongside the built-in ones. 证据：`docs/sources/configure/proxied-tools.md`
- **Developer**（documentation）：Build and test the server from source, use it as a Go library, and run it with Prometheus metrics and OpenTelemetry tracing. 证据：`docs/sources/developer/_index.md`
- **Build, test, and lint**（documentation）：Contributions are welcome. Open an issue or pull request on GitHub https://github.com/grafana/mcp-grafana . This project is written in Go. 证据：`docs/sources/developer/build-and-test.md`
- **Go SDK programmatic use**（documentation）：You can use the Grafana MCP server as a Go library to build custom MCP server contexts, for example with custom TLS or debug settings. The package is available on pkg.go.dev https://pkg.go.dev/github.com/grafana/mcp-grafana . 证据：`docs/sources/developer/go-sdk.md`
- **Observability metrics, tracing, and logs**（documentation）：Observability metrics, tracing, and logs 证据：`docs/sources/developer/observability-metrics-and-tracing.md`
- **Guides**（documentation）：Step-by-step use cases for the Grafana MCP server: query metrics and logs, search and inspect dashboards, manage alert rules, generate deeplinks, run panel queries, and use Grafana Incident and Sift. 证据：`docs/sources/guides/_index.md`
- **Generate deeplinks to Grafana**（documentation）：Use the Grafana MCP server so your AI assistant can generate correct deeplink URLs to dashboards, panels, and Grafana Explore. The server uses the Grafana API to build URLs so you get working links instead of guessed paths. 证据：`docs/sources/guides/generate-deeplinks-to-grafana.md`
- **Manage alert rules**（documentation）：Use the Grafana MCP server so your AI assistant can list and fetch alert rules, create or update rules, delete rules, and inspect notification policies and contact points. 证据：`docs/sources/guides/manage-alert-rules.md`
- **Query logs with Loki**（documentation）：Use the Grafana MCP server so your AI assistant can run LogQL log and metric queries against a Loki datasource in Grafana. You can also list label names and values and retrieve log patterns. 证据：`docs/sources/guides/query-logs-with-loki.md`
- **Query metrics with InfluxDB**（documentation）：Use the Grafana MCP server so your AI assistant can run InfluxDB queries against an InfluxDB datasource in Grafana. The server supports both InfluxQL v1.x and Flux v2.x ; the dialect is inferred from the datasource's configured version or can be set explicitly. 证据：`docs/sources/guides/query-metrics-with-influxdb.md`
- **Query metrics with Prometheus**（documentation）：Use the Grafana MCP server so your AI assistant can run PromQL queries against a Prometheus datasource in Grafana. You can run instant or range queries and discover metric and label names and values. 证据：`docs/sources/guides/query-metrics-with-prometheus.md`
- **Run a dashboard panel query**（documentation）：Use the Grafana MCP server to run a dashboard panel’s query the same query the panel uses with a custom time range and variable overrides. This is useful when you want the assistant to “run what this panel runs” without writing PromQL or LogQL yourself. 证据：`docs/sources/guides/run-a-dashboard-panel-query.md`
- **Search and inspect dashboards**（documentation）：Use the Grafana MCP server so your AI assistant can search for dashboards and inspect them summaries, panel queries, or specific properties without pulling full dashboard JSON. That keeps context window use under control. 证据：`docs/sources/guides/search-and-inspect-dashboards.md`
- **Use Grafana Incident and Sift**（documentation）：Use the Grafana MCP server so your AI assistant can work with Grafana Incident list, create, get, add activities and Sift list investigations, get analyses, find error patterns in logs, find slow requests . These features use Grafana basic roles: Viewer for read, Editor for write. 证据：`docs/sources/guides/use-grafana-incident-and-sift.md`
- **Introduction to the Grafana MCP server**（documentation）：Introduction to the Grafana MCP server 证据：`docs/sources/introduction.md`
- **Reference**（documentation）：Technical reference for the Grafana MCP server, including the MCP tools permission table. 证据：`docs/sources/reference/_index.md`
- **MCP tools reference**（documentation）：Use the table to confirm minimum Grafana RBAC permissions and scopes for each MCP tool. The sections after the table summarize RBAC patterns, optional categories, and a few operational notes. 证据：`docs/sources/reference/mcp-tools-table.md`
- **Set up the Grafana MCP server**（documentation）：Choose how you install and run the Grafana MCP server. Start with uvx for the least setup, or use Docker, a downloaded binary, or Helm when that fits your environment. Refer to Clients ../clients/ for client-specific steps and Client configuration examples client-configuration-examples/ for copy-paste MCP JSON debug, TLS, and more . 证据：`docs/sources/set-up/_index.md`
- **Install the binary**（documentation）：Install the Grafana MCP server by downloading a release binary or building from source. This gives you a single executable in your $PATH . 证据：`docs/sources/set-up/install-the-binary.md`
- **Troubleshooting**（documentation）：Diagnose common problems with the Grafana MCP server and Grafana compatibility. 证据：`docs/sources/troubleshooting/_index.md`
- **Server**（structured_config）：{ "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json", "name": "io.github.grafana/mcp-grafana", "description": "An MCP server giving access to Grafana dashboards, data and more.", "repository": { "url": "https://github.com/grafana/mcp-grafana", "source": "github" }, "version": "$VERSION", "icon": "assets/icon.png", "remotes": { "type": "streamable-http", "url": "https://mcp.grafana.com/mcp", "headers": { "name": "X-Grafana-URL", "description": "URL of your Grafana Cloud instance", "placeholder": "https:// .grafana.net", "isRequired": false, "isSecret": false } } , "packages": { "registryType": "oci", "identifier": "docker.io/grafana/mcp-grafana:$VERSION… 证据：`server.json`
- **Build stage**（source_file）：Build stage FROM --platform=$BUILDPLATFORM golang:1.26-alpine AS builder 证据：`Dockerfile.alpine`
- **Client Cache**（source_file）：package mcpgrafana ⋮---- import "context" "crypto/sha256" "fmt" "log/slog" "net/http" "net/url" "sort" "strings" "sync" "github.com/grafana/incident-go" "go.opentelemetry.io/otel" "go.opentelemetry.io/otel/attribute" "go.opentelemetry.io/otel/metric" "golang.org/x/sync/singleflight" ⋮---- "context" "crypto/sha256" "fmt" "log/slog" "net/http" "net/url" "sort" "strings" "sync" ⋮---- "github.com/grafana/incident-go" "go.opentelemetry.io/otel" "go.opentelemetry.io/otel/attribute" "go.opentelemetry.io/otel/metric" "golang.org/x/sync/singleflight" ⋮---- const clientCacheMeterName = "mcp-grafana" ⋮---- type clientCacheKey struct { url string apiKey string username string password string orgID int6… 证据：`client_cache.go`
- **Docker Compose**（source_file）：services: grafana: image: grafana/grafana:13.0.1@sha256:0f86bada30d65ef9d0183b90c1e2682ac92d53d95da8bed322b984ea78a4a73a environment: GF AUTH ANONYMOUS ENABLED: "false" GF INSTALL PLUGINS: grafana-clickhouse-datasource,grafana-opensearch-datasource,elasticsearch 12.5.4 GF LOG LEVEL: debug GF SERVER ROUTER LOGGING: "true" GF RENDERING SERVER URL: http://renderer:8081/render GF RENDERING CALLBACK URL: http://grafana:3000/ GF RENDERING RENDERER TOKEN: integration-test-renderer-token GF FEATURE TOGGLES ENABLE: provisioning,kubernetesDashboards ports: - 3000:3000/tcp volumes: - ./testdata/provisioning:/etc/grafana/provisioning - ./testdata/dashboards:/var/lib/grafana/dashboards - type: tmpfs tar… 证据：`docker-compose.yaml`
- **Mcpgrafana**（source_file）：package mcpgrafana ⋮---- import "context" "crypto/tls" "crypto/x509" "encoding/json" "fmt" "io" "log/slog" "net" "net/http" "net/textproto" "net/url" "os" "reflect" "runtime/debug" "strconv" "strings" "sync" "time" "github.com/go-openapi/runtime" openapiclient "github.com/go-openapi/runtime/client" "github.com/go-openapi/strfmt" "github.com/grafana/grafana-openapi-client-go/client" "github.com/grafana/incident-go" "github.com/mark3labs/mcp-go/server" "go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp" "golang.org/x/sync/singleflight" ⋮---- "context" "crypto/tls" "crypto/x509" "encoding/json" "fmt" "io" "log/slog" "net" "net/http" "net/textproto" "net/url" "os" "reflect" "runtime… 证据：`mcpgrafana.go`
- **Logs**（source_file）：package observability ⋮---- import "context" "errors" "fmt" "log/slog" "os" "runtime/debug" "go.opentelemetry.io/otel/exporters/otlp/otlplog/otlploggrpc" sdklog "go.opentelemetry.io/otel/sdk/log" sdkresource "go.opentelemetry.io/otel/sdk/resource" ⋮---- "context" "errors" "fmt" "log/slog" "os" "runtime/debug" ⋮---- "go.opentelemetry.io/otel/exporters/otlp/otlplog/otlploggrpc" sdklog "go.opentelemetry.io/otel/sdk/log" sdkresource "go.opentelemetry.io/otel/sdk/resource" ⋮---- func OTLPLogsEndpoint string ⋮---- type fanoutHandler struct { children slog.Handler } ⋮---- func NewFanoutHandler children ...slog.Handler slog.Handler ⋮---- func f fanoutHandler Enabled ctx context.Context, level slog.… 证据：`observability/logs.go`
- **Session**（source_file）：package mcpgrafana ⋮---- import "context" "fmt" "log/slog" "sync" "time" "github.com/mark3labs/mcp-go/mcp" "github.com/mark3labs/mcp-go/server" "go.opentelemetry.io/otel" "go.opentelemetry.io/otel/metric" ⋮---- "context" "fmt" "log/slog" "sync" "time" ⋮---- "github.com/mark3labs/mcp-go/mcp" "github.com/mark3labs/mcp-go/server" "go.opentelemetry.io/otel" "go.opentelemetry.io/otel/metric" ⋮---- const DefaultSessionTTL = 30 time.Minute sessionMeterName = "mcp-grafana" ⋮---- type sessionMetrics struct { activeSessions metric.Int64Gauge sessionsReaped metric.Int64Counter } ⋮---- func newSessionMetrics sessionMetrics ⋮---- type SessionState struct { lastActivity time.Time initOnce sync.Once proxi… 证据：`session.go`
- 其余 20 条证据见 `AI_CONTEXT_PACK.json` 或 `EVIDENCE_INDEX.json`。

## 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：`docs/docs.agent/README.md`, `README.md`, `internal/linter/jsonschema/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：`docs/docs.agent/README.md`, `README.md`, `internal/linter/jsonschema/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

下面内容用于强化 Repomix/AI Context Pack 主体。Human Manual 只提供阅读骨架；踩坑日志会被转成宿主 AI 必须遵守的工作约束。

## Human Manual 骨架

使用规则：这里只是项目阅读路线和显著性信号，不是事实权威。具体事实仍必须回到 repo evidence / Claim Graph。

宿主 AI 硬性规则：
- 不得把页标题、章节顺序、摘要或 importance 当作项目事实证据。
- 解释 Human Manual 骨架时，必须明确说它只是阅读路线/显著性信号。
- 能力、安装、兼容性、运行状态和风险判断必须引用 repo evidence、source path 或 Claim Graph。

- **项目概述**：importance `high`
  - source_paths: README.md, mcpgrafana.go, server.json
- **快速开始**：importance `high`
  - source_paths: docs/sources/set-up/install-with-uvx.md, docs/sources/set-up/install-with-docker.md, docs/sources/set-up/deploy-with-helm.md, docker-compose.yaml
- **安装与配置**：importance `high`
  - source_paths: docs/sources/configure/command-line-flags.md, docs/sources/configure/transports-and-addresses.md, docs/sources/set-up/client-configuration-examples.md, cmd/mcp-grafana/main.go
- **认证与授权**：importance `high`
  - source_paths: docs/sources/configure/authentication.md, docs/sources/configure/multi-organization-and-headers.md, client_cache.go, session.go
- **部署与运维**：importance `medium`
  - source_paths: Dockerfile, Dockerfile.alpine, docs/sources/set-up/deploy-with-helm.md, docs/sources/configure/server-tls-streamable-http.md, session_horizontal_scaling_test.go
- **仪表板工具**：importance `high`
  - source_paths: tools/dashboard.go, tools/dashboard_helpers.go, tools/navigation.go, tools/rendering.go
- **数据源工具**：importance `high`
  - source_paths: tools/datasources.go, tools/ds_query.go, tools/snowflake.go, tools/athena.go, tools/clickhouse.go
- **查询执行工具**：importance `high`
  - source_paths: tools/prometheus.go, tools/loki.go, tools/loki_backend.go, tools/influxdb.go, tools/graphite.go

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `7e42c9771c5ad145d4a9b7847cab8819d179f18a`
- inspected_files: `Dockerfile`, `README.md`, `docs/troubleshooting.md`, `docs/sources/introduction.md`, `docs/sources/_index.md`, `docs/docs.agent/context.md`, `docs/docs.agent/links.py`, `docs/docs.agent/rules.md`, `docs/docs.agent/README.md`, `docs/sources/set-up/install-with-uvx.md`, `docs/sources/set-up/deploy-with-helm.md`, `docs/sources/set-up/install-with-docker.md`, `docs/sources/set-up/client-configuration-examples.md`, `docs/sources/set-up/install-the-binary.md`, `docs/sources/set-up/_index.md`, `docs/sources/reference/mcp-tools-table.md`, `docs/sources/reference/_index.md`, `docs/sources/guides/use-grafana-incident-and-sift.md`, `docs/sources/guides/query-metrics-with-prometheus.md`, `docs/sources/guides/run-a-dashboard-panel-query.md`

宿主 AI 硬性规则：
- 没有 repo_clone_verified=true 时，不得声称已经读过源码。
- 没有 repo_inspection_verified=true 时，不得把 README/docs/package 文件判断写成事实。
- 没有 quick_start_verified=true 时，不得声称 Quick Start 已跑通。

## Doramagic Pitfall Constraints / 踩坑约束

这些规则来自 Doramagic 发现、验证或编译过程中的项目专属坑点。宿主 AI 必须把它们当作工作约束，而不是普通说明文字。

### Constraint 1: 来源证据：Insturctions

- 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：Insturctions
- 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_88b5b9a684e64beda991ee684f1f6daa | https://github.com/grafana/mcp-grafana/issues/908 | 来源类型 github_issue 暴露的待验证使用条件。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 2: 来源证据：Add tool to create Grafana short URLs (/goto/<uid>) via /api/short-urls

- 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：Add tool to create Grafana short URLs (/goto/<uid>) via /api/short-urls
- 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: 可能影响升级、迁移或版本选择。
- Evidence: community_evidence:github | cevd_9cf63ed3a3854737bbca0bbe18c11907 | https://github.com/grafana/mcp-grafana/issues/820 | 来源类型 github_issue 暴露的待验证使用条件。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 3: 来源证据：Feature Request: OAuth/SSO Authentication Support (Okta, Google, etc.)

- 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 Request: OAuth/SSO Authentication Support (Okta, Google, etc.)
- 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_ec3078db55a742068339d4ce3b6cc13f | https://github.com/grafana/mcp-grafana/issues/284 | 来源类型 github_issue 暴露的待验证使用条件。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 4: 来源证据：Feature request: pagination / chunking for query_loki_logs results

- 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 request: pagination / chunking for query_loki_logs results
- 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_cb6f3653732644d387da810f289708c4 | https://github.com/grafana/mcp-grafana/issues/761 | 来源讨论提到 windows 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 5: 来源证据：Security consideration: Prompt injection via dashboard data in MCP context

- 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：Security consideration: Prompt injection via dashboard data in MCP context
- 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_91b6f0b129764ca0b5d8cec3ee1497a7 | https://github.com/grafana/mcp-grafana/issues/680 | 来源讨论提到 api key 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 6: 来源证据：`get_panel_image` should support `orgId` when rendering panels

- 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：`get_panel_image` should support `orgId` when rendering panels
- 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_a05127eb498a4339b382b65279eef3de | https://github.com/grafana/mcp-grafana/issues/883 | 来源类型 github_issue 暴露的待验证使用条件。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 7: 能力判断依赖假设

- Trigger: README/documentation is current enough for a first validation pass.
- Host AI rule: 将假设转成下游验证清单。
- Why it matters: 假设不成立时，用户拿不到承诺的能力。
- Evidence: capability.assumptions | github_repo:907869862 | https://github.com/grafana/mcp-grafana | README/documentation is current enough for a first validation pass.
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 8: 维护活跃度未知

- Trigger: 未记录 last_activity_observed。
- Host AI rule: 补 GitHub 最近 commit、release、issue/PR 响应信号。
- Why it matters: 新项目、停更项目和活跃项目会被混在一起，推荐信任度下降。
- Evidence: evidence.maintainer_signals | github_repo:907869862 | https://github.com/grafana/mcp-grafana | last_activity_observed missing
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 9: 下游验证发现风险项

- Trigger: no_demo
- Host AI rule: 进入安全/权限治理复核队列。
- Why it matters: 下游已经要求复核，不能在页面中弱化。
- Evidence: downstream_validation.risk_items | github_repo:907869862 | https://github.com/grafana/mcp-grafana | no_demo; severity=medium
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 10: 存在评分风险

- Trigger: no_demo
- Host AI rule: 把风险写入边界卡，并确认是否需要人工复核。
- Why it matters: 风险会影响是否适合普通用户安装。
- Evidence: risks.scoring_risks | github_repo:907869862 | https://github.com/grafana/mcp-grafana | no_demo; severity=medium
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。
