# ray - 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 ray 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

- **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/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `.claude/skills/ray-dependencies/SKILL.md`, `.claude/skills/rebuild/SKILL.md` 等 Claim：`clm_0003` 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/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `.claude/skills/ray-dependencies/SKILL.md`, `.claude/skills/rebuild/SKILL.md` 等 Claim：`clm_0001` supported 0.86, `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：`.claude/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `doc/source/cluster/kubernetes/examples/distributed-checkpointing-with-gcsfuse.md`, `doc/source/cluster/kubernetes/examples/mnist-training-example.md` 等 Claim：`clm_0001` supported 0.86, `clm_0002` supported 0.86

## 怎么开始

- `curl -s -H "Authorization: Bearer $BUILDKITE_API_TOKEN" \` 证据：`.claude/skills/fetch-buildkite-logs/SKILL.md` Claim：`clm_0004` supported 0.86
- `pip install -c python/requirements_compiled.txt pre-commit && pre-commit install` 证据：`.claude/skills/lint/SKILL.md` Claim：`clm_0005` supported 0.86
- `curl -LO https://raw.githubusercontent.com/ray-project/kuberay/master/ray-operator/config/samples/pytorch-resnet-image-classifier/ray-job.pytorch-image-classifier.yaml` 证据：`doc/source/cluster/kubernetes/examples/distributed-checkpointing-with-gcsfuse.md` Claim：`clm_0006` unverified 0.25
- `curl -LO https://raw.githubusercontent.com/ray-project/kuberay/master/ray-operator/config/samples/pytorch-mnist/ray-job.pytorch-mnist.yaml` 证据：`doc/source/cluster/kubernetes/examples/mnist-training-example.md` Claim：`clm_0007` unverified 0.25
- `curl -LO https://raw.githubusercontent.com/ray-project/kuberay/v1.6.0/ray-operator/config/samples/ray-job.batch-inference.yaml` 证据：`doc/source/cluster/kubernetes/examples/rayjob-batch-inference-example.md` Claim：`clm_0008` unverified 0.25
- `curl -LO https://raw.githubusercontent.com/ray-project/kuberay/master/ray-operator/config/samples/pytorch-text-classifier/ray-job.pytorch-distributed-training.yaml` 证据：`doc/source/cluster/kubernetes/examples/rayjob-kueue-gang-scheduling.md` Claim：`clm_0009` unverified 0.25
- `curl -o ray-service.llm-serve.yaml https://raw.githubusercontent.com/ray-project/kuberay/master/ray-operator/config/samples/ray-service.llm-serve.yaml` 证据：`doc/source/cluster/kubernetes/examples/rayserve-llm-example.md` Claim：`clm_0010` unverified 0.25
- `curl --location 'http://localhost:8000/v1/chat/completions' --header 'Content-Type: application/json'` 证据：`doc/source/cluster/kubernetes/examples/rayserve-llm-example.md` Claim：`clm_0011` unverified 0.25
- `curl -LO https://raw.githubusercontent.com/ray-project/serve_config_examples/master/stable_diffusion/stable_diffusion_req.py` 证据：`doc/source/cluster/kubernetes/examples/stable-diffusion-rayservice.md` Claim：`clm_0012` unverified 0.25
- `curl -LO https://raw.githubusercontent.com/ray-project/serve_config_examples/master/text_summarizer/text_summarizer_req.py` 证据：`doc/source/cluster/kubernetes/examples/text-summarizer-rayservice.md` Claim：`clm_0013` unverified 0.25

## 继续前判断卡

- **当前建议**：仅建议沙盒试装
- **为什么**：项目存在安装命令、宿主配置或本地写入线索，不建议直接进入主力环境，应先在隔离环境试装。

### 30 秒判断

- **现在怎么做**：仅建议沙盒试装
- **最小安全下一步**：先跑 Prompt Preview；若仍要安装，只在隔离环境试装
- **先别相信**：真实输出质量不能在安装前相信。
- **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：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/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `.claude/skills/ray-dependencies/SKILL.md`, `.claude/skills/rebuild/SKILL.md` 等 Claim：`clm_0003` 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/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `.claude/skills/ray-dependencies/SKILL.md`, `.claude/skills/rebuild/SKILL.md` 等 Claim：`clm_0001` supported 0.86, `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：`.claude/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `doc/source/cluster/kubernetes/examples/distributed-checkpointing-with-gcsfuse.md`, `doc/source/cluster/kubernetes/examples/mnist-training-example.md` 等 Claim：`clm_0001` supported 0.86, `clm_0002` 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：`.claude/skills/fetch-buildkite-logs/SKILL.md` Claim：`clm_0004` supported 0.86

### 现在还不能相信

- **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/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `.claude/skills/ray-dependencies/SKILL.md`, `.claude/skills/rebuild/SKILL.md` 等
- **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。
- **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：`.claude/skills/fetch-buildkite-logs/SKILL.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：`.claude/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `doc/source/cluster/kubernetes/examples/distributed-checkpointing-with-gcsfuse.md`, `doc/source/cluster/kubernetes/examples/mnist-training-example.md` 等
- **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/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `.claude/skills/ray-dependencies/SKILL.md`, `.claude/skills/rebuild/SKILL.md` 等
- **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/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `doc/source/cluster/kubernetes/examples/distributed-checkpointing-with-gcsfuse.md`, `doc/source/cluster/kubernetes/examples/mnist-training-example.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。）
- **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。
- **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_0028` 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：`.claude/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `doc/source/cluster/kubernetes/examples/distributed-checkpointing-with-gcsfuse.md`, `doc/source/cluster/kubernetes/examples/mnist-training-example.md` 等 Claim：`clm_0029` 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/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `.claude/skills/ray-dependencies/SKILL.md`, `.claude/skills/rebuild/SKILL.md` 等 Claim：`clm_0001` supported 0.86, `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：`.claude/skills/fetch-buildkite-logs/SKILL.md`, `.claude/skills/lint/SKILL.md`, `doc/source/cluster/kubernetes/examples/distributed-checkpointing-with-gcsfuse.md`, `doc/source/cluster/kubernetes/examples/mnist-training-example.md` 等 Claim：`clm_0001` supported 0.86, `clm_0002` supported 0.86

### 上下文规模

- 文件总数：9026
- 重要文件覆盖：40/9026
- 证据索引条目：80
- 角色 / Skill 条目：6

### 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 ray 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 ray 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 ray 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 6 个角色 / Skill / 项目文档条目。

- **fetch-buildkite-logs**（skill）：Fetch and analyze Buildkite CI build logs for failures 激活提示：当用户任务与“fetch-buildkite-logs”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/fetch-buildkite-logs/SKILL.md`
- **lint**（skill）：Run linting and formatting checks on Ray code 激活提示：当用户任务与“lint”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/lint/SKILL.md`
- **ray-dependencies**（skill）：Manage Python dependencies in Ray — add/remove/upgrade packages, work with raydepsets lock files, debug dependency conflicts, and regenerate compiled requirements. Covers python/requirements , python/requirements/ , python/deplocks/ , and ci/raydepsets/configs/ .depsets.yaml . 激活提示：当用户任务与“ray-dependencies”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/ray-dependencies/SKILL.md`
- **rebuild**（skill）：Rebuild Ray from source — determines the right build mode based on what changed 激活提示：当用户任务与“rebuild”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/rebuild/SKILL.md`
- **rst-to-myst**（skill）：Convert Ray documentation pages from reStructuredText .rst to MyST Markdown .md . Use when migrating existing files under doc/source/ to MyST, finishing a partial MyST migration of a directory, or when asked to convert/migrate a doc page to markdown. Covers the RST-to-MyST directive mapping, label and cross-reference preservation, sphinx-design tabs/dropdowns, doctest/testcode handling, the doc/BUILD.bazel doctest e… 激活提示：当用户任务与“rst-to-myst”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：`doc/.claude/skills/rst-to-myst/SKILL.md`
- **sphinx-fix**（skill）：Diagnose a failing Ray Sphinx / Read the Docs documentation build. Parses the Sphinx warning stream an RtD build log, a local build, or pasted text , classifies each warning against a rules table, and proposes the canonical fix in severity-tier order. Detects a hard-broken build, segregates known-benign suppressed classes, and lists every unclassified warning. Use when a docs/readthedocs.com:anyscale-ray check fails… 激活提示：当用户任务与“sphinx-fix”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：`doc/.claude/skills/sphinx-fix/SKILL.md`

## 证据索引

- 共索引 80 条证据。

- **Buildkite pipelines**（documentation）：This directory contains the Buildkite pipeline definitions for Ray CI, plus the rules that decide which tests run on a given change. 证据：`.buildkite/README.md`
- **Ray**（documentation）：Ray is a unified framework for scaling AI and Python applications. 证据：`.claude/CLAUDE.md`
- **Ray Documentation**（documentation）：Repository for documentation of the Ray project, hosted at docs.ray.io https://docs.ray.io . 证据：`doc/README.md`
- **Readme**（documentation）：Overview of how the ray images are built: 证据：`docker/README.md`
- **Instructions on ray-java test**（documentation）：1. Install necessary executables - java and javac is needed to run ray-java tests, and users need to make sure they're accessible in $PATH - You could check whether they're installed by which java and which javac - Install java with sudo apt install openjdk-11-jre -y - Install javac with sudo apt install openjdk-11-jdk -y - java-11 is the version we use on CI 证据：`java/README.md`
- **Readme**（documentation）：/ .md -- 证据：`.claude/agents/README.md`
- **raydepsets**（documentation）：A dependency lock file management tool for Ray CI pipelines. It maintains consistency relationships among lock files — ensuring that when a dependency is updated, all related lock files are regenerated together in the correct order. Built on top of uv pip compile https://docs.astral.sh/uv/pip/compile/ , it generates reproducible, hash-verified lock files across multiple Python versions, platforms, and CUDA variants. 证据：`ci/raydepsets/README.md`
- **Ray Documentation**（documentation）：Sphinx documentation built by Read the Docs and served at docs.ray.io . Build pipeline: .buildkite/doc.rayci.yml . 证据：`doc/.claude/CLAUDE.md`
- **Read the Docs redirects for docs.ray.io**（documentation）：Read the Docs redirects for docs.ray.io 证据：`doc/redirects/README.md`
- **Video analysis inference pipeline with Ray Serve**（documentation）：Video analysis inference pipeline with Ray Serve 证据：`doc/source/serve/tutorials/video-analysis/README.md`
- **Scaling Batch Inference with Ray Data**（documentation）：Scaling Batch Inference with Ray Data 证据：`doc/source/templates/01_batch_inference/README.md`
- **Scaling Many Model Training with Ray Tune**（documentation）：Scaling Many Model Training with Ray Tune 证据：`doc/source/templates/02_many_model_training/README.md`
- **Serving a Stable Diffusion Model with Ray Serve**（documentation）：Serving a Stable Diffusion Model with Ray Serve 证据：`doc/source/templates/03_serving_stable_diffusion/README.md`
- **Fine-tuning Llama-2 series models with Deepspeed, Accelerate, and Ray Train TorchTrainer**（documentation）：Fine-tuning Llama-2 series models with Deepspeed, Accelerate, and Ray Train TorchTrainer Template Specification Description ---------------------- ----------- Summary This template, demonstrates how to perform fine-tuning full parameter or LoRA for Llama-2 series models 7B, 13B, and 70B using TorchTrainer with the DeepSpeed ZeRO-3 strategy. Time to Run 1 epoch 3.5M tokens training wall-clock time: ~14 min. for 7B, ~26 min. for 13B, and ~190 min. for 70B see the setup details below Minimum Compute Requirements 16xg5.4xlarge for worker nodes for 7B model, 4xg5.12xlarge nodes for 13B model, and 4xg5.48xlarge or 2xp4de.24xlarge nodes for 70B Cluster Environment This template uses a Docker image… 证据：`doc/source/templates/04_finetuning_llms_with_deepspeed/README.md`
- **DreamBooth fine-tuning of Stable Diffusion with Ray Train**（documentation）：DreamBooth fine-tuning of Stable Diffusion with Ray Train 证据：`doc/source/templates/05_dreambooth_finetuning/README.md`
- **Ray Starter Templates**（documentation）：These templates are a set of minimal examples that are quick and easy to run and customize. 证据：`doc/source/templates/README.md`
- **Train a GPT-2 model with Ray Train JaxTrainer**（documentation）：Train a GPT-2 model with Ray Train JaxTrainer 证据：`doc/source/train/examples/jax/intro_to_jax_trainer/README.md`
- **RL Post-Training using Hugging Face TRL with GRPO**（documentation）：RL Post-Training using Hugging Face TRL with GRPO 证据：`doc/source/train/examples/transformers/transformer_reinforcement_learning/README.md`
- **DEPRECATED -- Please use rayproject/ray-ml https://hub.docker.com/repository/docker/rayproject/ray-ml**（documentation）：DEPRECATED -- Please use rayproject/ray-ml https://hub.docker.com/repository/docker/rayproject/ray-ml 证据：`docker/autoscaler/README.md`
- **About**（documentation）：About This is an internal image, the rayproject/ray https://hub.docker.com/repository/docker/rayproject/ray or rayproject/ray-ml https://hub.docker.com/repository/docker/rayproject/ray-ml should be used! 证据：`docker/base-deps/README.md`
- **About**（documentation）：About This image is an extension of the rayproject/ray https://hub.docker.com/repository/docker/rayproject/ray image. It includes all extended requirements of RLlib , Serve and Tune . It is a well-provisioned starting point for trying out the Ray ecosystem. Find the Dockerfile here. https://github.com/ray-project/ray/blob/master/docker/ray-ml/Dockerfile 证据：`docker/ray-ml/README.md`
- **Tags**（documentation）：Official container images for Ray https://github.com/ray-project/ray . These images contains working Python virtual environments and required dependencies to run launche Ray nodes and form Ray clusters. everything needed to get started with running Ray. One can use the images for local development, or launch clusters with Ray VM launcher vm-launcher , KubeRay kuberay , or running them on Anyscale anyscale . 证据：`docker/ray/README.md`
- **Updating this Lambda Function**（documentation）：3. Head to the AWS Management console & select the DockerTagLatest function. Select Upload from , then .zip file and then select the zip file created in Step 2. 证据：`docker/retag-lambda/README.md`
- **Common Utilities Shared Across the Libraries**（documentation）：Common Utilities Shared Across the Libraries 证据：`python/ray/_common/README.md`
- **How to pull upstream changes into the vendored cloudpickle**（documentation）：How to pull upstream changes into the vendored cloudpickle 证据：`python/ray/cloudpickle/README.md`
- **Ray Data**（documentation）：Ray Data Key Modules Gotchas 证据：`python/ray/data/.claude/CLAUDE.md`
- **Readme**（documentation）： 证据：`python/ray/data/.claude/rules/README.md`
- **Video Processing Example**（documentation）：This folder contains a self-contained example that demonstrates how Ray Data can prepare video inputs before they are passed to a multimodal model. The implementation lives in video processor.py ; it focuses on being a small, re-usable utility that you can compose inside map batches or call directly from an async workflow. 证据：`python/ray/data/examples/data/video_processing/README.md`
- **Multi-Turn LLM Benchmark**（documentation）：A benchmark tool for OpenAI-compatible LLM inference servers that supports multi-turn conversations with configurable prefix cache hit rates, input/output sequence lengths, and cross-session prefix sharing. 证据：`python/ray/llm/_internal/serve/benchmark/README.md`
- **SGLang on Ray Serve LLM**（documentation）：This directory contains example scripts for using SGLang with Ray Serve LLM. 证据：`python/ray/llm/examples/sglang/readme.md`
- **Ray Serve**（documentation）：Ray Serve Key Modules Gotchas 证据：`python/ray/serve/.claude/CLAUDE.md`
- **Readme**（documentation）： 证据：`python/ray/serve/.claude/rules/README.md`
- **Ray Train**（documentation）：Ray Train Key Modules Gotchas 证据：`python/ray/train/.claude/CLAUDE.md`
- **Readme**（documentation）： 证据：`python/ray/train/.claude/rules/README.md`
- **Ray Train Circular Import Linter**（documentation）：Ray Train functionality is overrided or "patched" by functionality from other directories. For instance, Ray Train is patched by functionality from Ray Train v2 when RAY TRAIN V2 ENABLED=1 , making Ray Train dependent on Ray Train v2. In turn, the patching directory often imports functionality from the "base" Ray Train directory ray/python/ray/train , resulting in a circular dependency. The Ray Train Circular Import Linter takes a patching directory, patch dir , and detects circular imports between it and the base Ray Train directory- displaying violations to users via pre-commit. 证据：`python/ray/train/lint/README.md`
- **Ray Tune**（documentation）：Ray Tune Key Modules Gotchas 证据：`python/ray/tune/.claude/CLAUDE.md`
- **Readme**（documentation）： 证据：`python/ray/tune/.claude/rules/README.md`
- **Ray Scalability Envelope**（documentation）：NOTE : the Ray scalability benchmarks are in the process of being refreshed. If you have questions about a specific workload or limit, please get in touch by filing a GitHub issue https://github.com/ray-project/ray/issues . 证据：`release/benchmarks/README.md`
- **Shared Profiling Module**（documentation）：Shared profiling and monitoring infrastructure for Ray Data benchmarks. Used by benchmarks under release/nightly tests/dataset/ e.g. image embedding from jsonl . 证据：`release/nightly_tests/dataset/profiling/README.md`
- **Profile Analysis Scripts**（documentation）：CLI tools for analyzing profiling output after a benchmark run. These operate on standard formats speedscope JSON, collapsed stacks and don't depend on the profiling module -- they can be used standalone. 证据：`release/nightly_tests/dataset/profiling/analysis/README.md`
- **Build Tools**（documentation）：Docker image build tooling for Ray Data benchmarks. 证据：`release/nightly_tests/dataset/profiling/build/README.md`
- **RLlib**（documentation）：RLlib Key Modules Gotchas 证据：`rllib/.claude/CLAUDE.md`
- **Readme**（documentation）： 证据：`rllib/.claude/rules/README.md`
- **Asynchronous Proximal Policy Optimization APPO**（documentation）：Asynchronous Proximal Policy Optimization APPO 证据：`rllib/algorithms/appo/README.md`
- **Conservative Q-Learning CQL**（documentation）：CQL https://arxiv.org/abs/2006.04779 is an offline RL algorithm that mitigates the overestimation of Q-values outside the dataset distribution via convservative critic estimates. CQL does this by adding a simple Q regularizer loss to the standard Belman update loss. This ensures that the critic does not output overly-optimistic Q-values and can be added on top of any off-policy Q-learning algorithm in this case, we use SAC . 证据：`rllib/algorithms/cql/README.md`
- **Deep Q Networks DQN**（documentation）：Code in this package is adapted from https://github.com/openai/baselines/tree/master/baselines/deepq. 证据：`rllib/algorithms/dqn/README.md`
- **DreamerV3**（documentation）：! DreamerV3 ../../../doc/source/rllib/images/dreamerv3/dreamerv3.png 证据：`rllib/algorithms/dreamerv3/README.md`
- **Proximal Policy Optimization PPO**（documentation）：PPO https://arxiv.org/abs/1707.06347 is a model-free on-policy RL algorithm that works well for both discrete and continuous action space environments. PPO utilizes an actor-critic framework, where there are two networks, an actor policy network and critic network value function . 证据：`rllib/algorithms/ppo/README.md`
- **Soft Actor Critic SAC**（documentation）：SAC https://arxiv.org/abs/1801.01290 is a SOTA model-free off-policy RL algorithm that performs remarkably well on continuous-control domains. SAC employs an actor-critic framework and combats high sample complexity and training stability via learning based on a maximum-entropy framework. Unlike the standard RL objective which aims to maximize sum of reward into the future, SAC seeks to optimize sum of rewards as well as expected entropy over the current policy. In addition to optimizing over an actor and critic with entropy-based objectives, SAC also optimizes for the entropy coeffcient. 证据：`rllib/algorithms/sac/README.md`
- **TQC Truncated Quantile Critics**（documentation）：TQC is an extension of SAC Soft Actor-Critic that uses distributional reinforcement learning with quantile regression to control overestimation bias in the Q-function. 证据：`rllib/algorithms/tqc/README.md`
- **Footsies Environment**（documentation）：This environment implementation is based on the FootsiesGym project https://github.com/chasemcd/FootsiesGym , specifically the version as of July 28, 2025 . 证据：`rllib/examples/envs/classes/multi_agent/footsies/README.md`
- **C++ Core Runtime**（documentation）：C++ Core Runtime Key Modules Gotchas 证据：`src/ray/.claude/CLAUDE.md`
- **Readme**（documentation）： 证据：`src/ray/.claude/rules/README.md`
- **Pubsub module**（documentation）：This doc has last been updated on Aug 19, 2025. This doc should be updated as the implementation changes. 证据：`src/ray/pubsub/README.md`
- **Package**（package_manifest）：{ "name": "ray-dashboard-client", "version": "1.0.0", "private": true, "dependencies": { "@emotion/react": "^11.11.3", "@emotion/styled": "^11.11.0", "@mui/icons-material": "^5.15.5", "@mui/material": "^5.15.5", "@reduxjs/toolkit": "^1.3.1", "@types/jest": "^27.5.2", "@types/lodash": "^4.14.161", "@types/node": "13.9.5", "@types/react-redux": "^7.1.7", "@types/react-window": "^1.8.2", "axios": "^0.21.1", "copy-to-clipboard": "^3.3.2", "dayjs": "^1.9.4", "js-yaml": "^4.1.0", "lodash": "^4.17.20", "lowlight": "^2.9.0", "react": "^18.3.0", "react-dom": "^18.3.0", "react-icons": "^4.7.1", "react-router-dom": "^6.4.3", "react-scripts": "^5.0.1", "react-window": "^1.8.5", "swr": "^2.1.0", "typefa… 证据：`python/ray/dashboard/client/package.json`
- **Fetch Buildkite Logs**（skill_instruction）：Prerequisites - BUILDKITE API TOKEN must be set in the environment typically ~/.bashrc - If not configured, direct user to doc/source/ray-contribute/agent-development.md for setup 证据：`.claude/skills/fetch-buildkite-logs/SKILL.md`
- **Lint Modified Files**（skill_instruction）：Run pre-commit on the files you changed: 证据：`.claude/skills/lint/SKILL.md`
- **Ray Dependencies**（skill_instruction）：Expert skill for managing Python dependencies across the Ray repository: the monorepo requirements compiled .txt lock files, the raydepsets DAG-based lock file manager, modular python/requirements/ source files, and Docker image dependency chains. 证据：`.claude/skills/ray-dependencies/SKILL.md`
- **Rebuild Ray**（skill_instruction）：Canonical build docs: doc/source/ray-contribute/development.md Use the user's configured Python from CLAUDE.local.md, or fall back to which python . Update this skill if any changes are detected in development.rst. 证据：`.claude/skills/rebuild/SKILL.md`
- **Convert RST to MyST Markdown**（skill_instruction）：MyST Markdown is the standard for new Ray doc pages — doc/.claude/CLAUDE.md declares it, and a lint check rejects newly-added .rst . This skill converts an existing .rst page or a batch to MyST .md faithfully : format only, preserving the rendered HTML and any test coverage. 证据：`doc/.claude/skills/rst-to-myst/SKILL.md`
- 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：`.buildkite/README.md`, `.claude/CLAUDE.md`, `doc/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：`.buildkite/README.md`, `.claude/CLAUDE.md`, `doc/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。

- **Ray 概览与系统架构**：importance `high`
  - source_paths: README.rst, python/ray/__init__.py, doc/source/index.rst, doc/source/ray-overview/ray-libraries.rst
- **Ray Core:任务、Actor 与对象**：importance `high`
  - source_paths: python/ray/_private/worker.py, python/ray/_private/runtime_env/runtime_env.py, python/ray/_raylet.pyx, python/ray/dag/compiled_dag_node.py, python/ray/experimental/rdt/rdt_manager.py
- **AI 库:Data、Train、Tune、RLlib、Serve、LLM**：importance `high`
  - source_paths: python/ray/data/dataset.py, python/ray/train/v2/_internal/execution/controller/controller.py, python/ray/tune/tune.py, python/ray/rllib, python/ray/serve/api.py
- **集群部署、Autoscaler 与可观测性**：importance `high`
  - source_paths: python/ray/autoscaler/v2/autoscaler.py, python/ray/dashboard/dashboard.py, python/ray/dashboard/modules/job/job_manager.py, python/ray/dashboard/modules/metrics/default_impl.py, python/ray/_private/authentication/authentication_token_setup.py

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `8946afb99cbc62c78faf4cd73b8232cf509e7ef5`
- inspected_files: `README.rst`, `pyproject.toml`, `src/ray/.claude/CLAUDE.md`, `src/ray/.claude/rules/README.md`, `src/ray/.claude/rules/cpp-style.md`, `src/ray/.cursor/BUGBOT.md`, `src/ray/design_docs/id_specification.md`, `src/ray/pubsub/README.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: 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 | https://github.com/ray-project/ray | 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 2: 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 | https://github.com/ray-project/ray | 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。

- Trigger: no_demo
- Evidence: downstream_validation.risk_items | https://github.com/ray-project/ray | 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 4: 存在评分风险

- Trigger: no_demo
- Why it matters: 风险会影响是否适合普通用户安装。
- Evidence: risks.scoring_risks | https://github.com/ray-project/ray | 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 5: issue/PR 响应质量未知

- Trigger: issue_or_pr_quality=unknown。
- Host AI rule: 抽样最近 issue/PR，判断是否长期无人处理。
- Why it matters: 用户无法判断遇到问题后是否有人维护。
- Evidence: evidence.maintainer_signals | https://github.com/ray-project/ray | issue_or_pr_quality=unknown
- 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: 发布节奏不明确

- Trigger: release_recency=unknown。
- Host AI rule: 确认最近 release/tag 和 README 安装命令是否一致。
- Why it matters: 安装命令和文档可能落后于代码，用户踩坑概率升高。
- Evidence: evidence.maintainer_signals | https://github.com/ray-project/ray | release_recency=unknown
- 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。