# openspace - 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 openspace 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_0003` 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：`gdpval_bench/skills/audio-track-production/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md` 等 Claim：`clm_0004` 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：`gdpval_bench/skills/audio-track-production/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md` 等 Claim：`clm_0001` 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_0002` supported 0.86

## 怎么开始

- `git clone https://github.com/HKUDS/OpenSpace.git && cd OpenSpace` 证据：`README.md` Claim：`clm_0005` supported 0.86
- `pip install -e .` 证据：`README.md` Claim：`clm_0006` supported 0.86
- `npm install        # only needed once` 证据：`README.md` Claim：`clm_0007` supported 0.86

## 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：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_0003` 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：`gdpval_bench/skills/audio-track-production/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md` 等 Claim：`clm_0004` 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：`gdpval_bench/skills/audio-track-production/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md` 等 Claim：`clm_0001` 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_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：`README.md` Claim：`clm_0005` 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：`gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced/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：`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：`gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced/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：`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**：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_0008` 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_0009` 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：`gdpval_bench/skills/audio-track-production/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`, `gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md` 等 Claim：`clm_0001` 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_0002` supported 0.86

### 上下文规模

- 文件总数：837
- 重要文件覆盖：40/837
- 证据索引条目：180
- 角色 / Skill 条目：260

### 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 openspace 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 openspace 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 openspace 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 260 个角色 / Skill / 项目文档条目。

- **adaptive-stem-alignment**（skill）：Incremental audio production with duration mismatch handling, adaptive stem extension, and pre-mix alignment verification 激活提示：当用户任务与“adaptive-stem-alignment”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：`gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md`
- **diagnostic-stem-delivery**（skill）：Audio production with diagnostic analysis, timecode parsing from documents, and verified export workflow 激活提示：当用户任务与“diagnostic-stem-delivery”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：`gdpval_bench/skills/audio-track-production-enhanced-enhanced-enhanced/SKILL.md`
- **aligned-stem-workflow**（skill）：Incremental audio production with duration alignment handling, per-stem verification, and adaptive extension strategies 激活提示：当用户任务与“aligned-stem-workflow”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：`gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`
- **incremental-audio-workflow**（skill）：Step-by-step audio production with per-stem verification, timing alignment, and incremental quality gates 激活提示：当用户任务与“incremental-audio-workflow”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：`gdpval_bench/skills/audio-track-production-enhanced/SKILL.md`
- **audio-track-production**（skill）：End-to-end audio production workflow with stems, effects, archiving, and verification 激活提示：当用户任务与“audio-track-production”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：`gdpval_bench/skills/audio-track-production/SKILL.md`
- **cascade-fail-recovery**（skill）：Handle cascading data retrieval tool failures by falling back to embedded knowledge generation 激活提示：当用户任务与“cascade-fail-recovery”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：`gdpval_bench/skills/cascade-fail-recovery/SKILL.md`
- **code-exec-fallback-266cba**（skill）：Fallback workflow for reliable code execution when sandbox fails repeatedly 激活提示：当用户任务与“code-exec-fallback-266cba”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：`gdpval_bench/skills/code-exec-fallback-266cba/SKILL.md`
- **code-exec-fallback**（skill）：Fallback pattern for executing Python code when execute code sandbox fails 激活提示：当用户任务与“code-exec-fallback”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：`gdpval_bench/skills/code-exec-fallback/SKILL.md`
- **code-execution-fallback-e81068**（skill）：Fallback workflow for executing Python code when execute code sandbox fails repeatedly 激活提示：当用户任务与“code-execution-fallback-e81068”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：`gdpval_bench/skills/code-execution-fallback-e81068/SKILL.md`
- **code-execution-fallback**（skill）：Handle code execution failures with fallback strategies and anchored workspace paths 激活提示：当用户任务与“code-execution-fallback”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：`gdpval_bench/skills/code-execution-fallback/SKILL.md`
- **comprehensive-soap-note**（skill）：Generate complete medical SOAP notes with all four sections in a single comprehensive file write operation 激活提示：当用户任务与“comprehensive-soap-note”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：`gdpval_bench/skills/create-soap-note-merged/SKILL.md`
- **create-soap-note**（skill）：Generate structured medical SOAP notes with all required sections in a single comprehensive file write. 激活提示：当用户任务与“create-soap-note”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：`gdpval_bench/skills/create-soap-note/SKILL.md`
- **debug-sandbox-execution**（skill）：Debug Python code execution failures by capturing partial traces, isolating failing functions, and incrementally verifying outputs 激活提示：当用户任务与“debug-sandbox-execution”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：`gdpval_bench/skills/debug-sandbox-execution/SKILL.md`
- **delegated-file-edit-verification**（skill）：Verify delegated file-edit results directly in the target workspace before accepting success, and rerun with absolute paths if the delegate's report conflicts with reality. 激活提示：当用户任务与“delegated-file-edit-verification”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：`gdpval_bench/skills/delegated-file-edit-verification/SKILL.md`
- **deliverable-completion**（skill）：Clarifies that file creation tasks are complete when the deliverable is successfully written—no submission step required 激活提示：当用户任务与“deliverable-completion”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：`gdpval_bench/skills/deliverable-completion/SKILL.md`
- **detect-zero-iteration-failure**（skill）：Identify and classify agent failures occurring before execution starts 激活提示：当用户任务与“detect-zero-iteration-failure”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：`gdpval_bench/skills/detect-zero-iteration-failure/SKILL.md`
- **detect-zero-iteration-failures**（skill）：Identify and handle pre-execution agent failures occurring before any iterations or tool usage 激活提示：当用户任务与“detect-zero-iteration-failures”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：`gdpval_bench/skills/detect-zero-iteration-failures/SKILL.md`
- **direct-reportlab-pdf-generation**（skill）：Generate complex multi-page PDFs by running reportlab Python code directly via run shell when shell agent fails on document creation 激活提示：当用户任务与“direct-reportlab-pdf-generation”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：`gdpval_bench/skills/direct-reportlab-pdf-generation/SKILL.md`
- **document-gen-resilient-workflow**（skill）：Multi-engine document generation with cascading PDF fallbacks and robust Unicode handling 激活提示：当用户任务与“document-gen-resilient-workflow”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：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-2794b4/SKILL.md`
- **doc-gen-unicode-diagnostic**（skill）：Systematic document generation with unicode sanitization, engine fallback chain, and explicit error diagnosis 激活提示：当用户任务与“doc-gen-unicode-diagnostic”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：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-3d3a9a/SKILL.md`
- **document-gen-resilient-multiformat**（skill）：Resilient multi-format document generation with environment checks, auto-sanitization, and fallback engines 激活提示：当用户任务与“document-gen-resilient-multiformat”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：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-618c9b/SKILL.md`
- **document-gen-resilient**（skill）：Multi-path document generation with tool checks, Unicode handling, and Python fallbacks 激活提示：当用户任务与“document-gen-resilient”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：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-96865f/SKILL.md`
- **document-gen-dual-backend**（skill）：Document generation with direct pandoc/ReportLab execution lightweight default and optional shell agent fallback for complex scenarios 激活提示：当用户任务与“document-gen-dual-backend”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：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-9f3b1f/SKILL.md`
- **document-gen-unicode-safe**（skill）：Unicode-safe fallback workflow for multi-format document generation with character sanitization 激活提示：当用户任务与“document-gen-unicode-safe”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：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-c5d426/SKILL.md`
- **document-gen-resilient-workflow**（skill）：Robust document generation with engine fallback chain, error diagnostics, and Python alternatives 激活提示：当用户任务与“document-gen-resilient-workflow”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：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced/SKILL.md`
- **resilient-document-pipeline**（skill）：Unified document generation with tool failure detection, domain knowledge fallback, progressive conversion, and Unicode safety 激活提示：当用户任务与“resilient-document-pipeline”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：`gdpval_bench/skills/document-gen-fallback-enhanced-merged/SKILL.md`
- **document-gen-unicode-safe**（skill）：Unicode-safe fallback workflow for multi-format document generation with character sanitization and explicit error capture 激活提示：当用户任务与“document-gen-unicode-safe”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：`gdpval_bench/skills/document-gen-fallback-enhanced/SKILL.md`
- **document-gen-fallback**（skill）：Fallback workflow for multi-format document generation when shell agent encounters errors 激活提示：当用户任务与“document-gen-fallback”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：`gdpval_bench/skills/document-gen-fallback/SKILL.md`
- **docx-read-fallback**（skill）：Use run shell with python-docx as reliable fallback when read file fails on .docx files 激活提示：当用户任务与“docx-read-fallback”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：`gdpval_bench/skills/docx-read-fallback/SKILL.md`
- **docx-shell-extract**（skill）：Extract text from DOCX files using shell commands when python-docx is unavailable 激活提示：当用户任务与“docx-shell-extract”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：`gdpval_bench/skills/docx-shell-extract/SKILL.md`
- **docx-dual-parse**（skill）：Extract text from DOCX files using shell or Python zipfile, with environment-aware fallback 激活提示：当用户任务与“docx-dual-parse”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：`gdpval_bench/skills/docx-shell-parse-enhanced-4bba79/SKILL.md`
- **docx-parse-resilient**（skill）：Extract text from DOCX files with shell-primary approach and Python zipfile fallback for maximum reliability 激活提示：当用户任务与“docx-parse-resilient”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：`gdpval_bench/skills/docx-shell-parse-enhanced/SKILL.md`
- **docx-shell-parse**（skill）：Extract text from DOCX files using shell commands when python-docx is unavailable 激活提示：当用户任务与“docx-shell-parse”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：`gdpval_bench/skills/docx-shell-parse/SKILL.md`
- **docx-shell-workaround**（skill）：Handle docx files using shell-based XML extraction and python-docx via run shell when standard tools fail 激活提示：当用户任务与“docx-shell-workaround”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：`gdpval_bench/skills/docx-shell-workaround/SKILL.md`
- **excel-creation-fallback**（skill）：Use shell agent as fallback when execute code sandbox fails for Excel file operations 激活提示：当用户任务与“excel-creation-fallback”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：`gdpval_bench/skills/excel-creation-fallback/SKILL.md`
- **excel-debug-extraction**（skill）：Iteratively debug Excel structure with exploratory scripts before writing extraction logic 激活提示：当用户任务与“excel-debug-extraction”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：`gdpval_bench/skills/excel-debug-extraction/SKILL.md`
- **excel-heredoc-fallback**（skill）：Create Excel files with openpyxl by falling back to shell-based Python heredoc when sandbox execution fails 激活提示：当用户任务与“excel-heredoc-fallback”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：`gdpval_bench/skills/excel-heredoc-fallback/SKILL.md`
- **excel-heredoc-workaround**（skill）：Create Excel files with openpyxl by falling back to run shell with inline Python heredoc when execute code sandbox fails 激活提示：当用户任务与“excel-heredoc-workaround”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：`gdpval_bench/skills/excel-heredoc-workaround/SKILL.md`
- **excel-subtotal-audit**（skill）：Audit generated Excel reports by enumerating populated output rows, exposing formulas, and checking that each subtotal range includes all visible detail rows. 激活提示：当用户任务与“excel-subtotal-audit”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：`gdpval_bench/skills/excel-subtotal-audit/SKILL.md`
- **excel-unmerge-before-write**（skill）：Unmerge merged cells in openpyxl worksheets before writing values to avoid AttributeError 激活提示：当用户任务与“excel-unmerge-before-write”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：`gdpval_bench/skills/excel-unmerge-before-write/SKILL.md`
- **execute-code-fallback**（skill）：Fallback from execute code sandbox to run shell when e2b sandbox is unavailable 激活提示：当用户任务与“execute-code-fallback”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：`gdpval_bench/skills/execute-code-fallback/SKILL.md`
- **fallback-code-execution**（skill）：Fallback workflow for running code via file write and shell when sandbox execution fails 激活提示：当用户任务与“fallback-code-execution”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：`gdpval_bench/skills/fallback-code-execution/SKILL.md`
- **fallback-file-creation**（skill）：Fallback pattern for reliable file creation when code execution tools fail 激活提示：当用户任务与“fallback-file-creation”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：`gdpval_bench/skills/fallback-file-creation/SKILL.md`
- **fallback-file-ops**（skill）：Use run shell as fallback when execute code sandbox or shell agent fail on filesystem operations 激活提示：当用户任务与“fallback-file-ops”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：`gdpval_bench/skills/fallback-file-ops/SKILL.md`
- **fallback-python-execution**（skill）：Reliable Python execution workflow when execute code sandbox or shell agent fail 激活提示：当用户任务与“fallback-python-execution”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：`gdpval_bench/skills/fallback-python-execution/SKILL.md`
- **fallback-python-shell**（skill）：Use run shell with Python heredoc when execute code sandbox or read file fail 激活提示：当用户任务与“fallback-python-shell”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：`gdpval_bench/skills/fallback-python-shell/SKILL.md`
- **fallback-script-execution**（skill）：Two-step script execution workflow for debugging when shell agent and execute code sandbox consistently fail 激活提示：当用户任务与“fallback-script-execution”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：`gdpval_bench/skills/fallback-script-execution/SKILL.md`
- **ffmpeg-encoder-check-4855c0**（skill）：Check available FFmpeg encoders before writing encoding scripts to avoid library version mismatches 激活提示：当用户任务与“ffmpeg-encoder-check-4855c0”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：`gdpval_bench/skills/ffmpeg-encoder-check-4855c0/SKILL.md`
- **ffmpeg-encoder-check**（skill）：Check FFmpeg encoder availability before video encoding to avoid library mismatches 激活提示：当用户任务与“ffmpeg-encoder-check”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：`gdpval_bench/skills/ffmpeg-encoder-check/SKILL.md`
- **ffmpeg-graceful-degradation**（skill）：Graceful degradation workflow for ffmpeg encoding failures with progressive fallback strategies 激活提示：当用户任务与“ffmpeg-graceful-degradation”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：`gdpval_bench/skills/ffmpeg-graceful-degradation/SKILL.md`
- **file-locate-verify**（skill）：Systematic workflow to resolve FileNotFoundError by locating and verifying file paths 激活提示：当用户任务与“file-locate-verify”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：`gdpval_bench/skills/file-locate-verify/SKILL.md`
- **finalize-zip-deliverable**（skill）：Ensures codebase deliverables are properly archived into a ZIP file before marking task complete 激活提示：当用户任务与“finalize-zip-deliverable”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：`gdpval_bench/skills/finalize-zip-deliverable/SKILL.md`
- **graceful-tool-failure**（skill）：Report tool failures transparently and attempt minimal viable output with disclaimers 激活提示：当用户任务与“graceful-tool-failure”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：`gdpval_bench/skills/graceful-tool-failure/SKILL.md`
- **heredoc-file-write**（skill）：Use shell heredoc syntax as a workaround when write file fails on complex TypeScript/JSX/code files with encoding or escaping issues 激活提示：当用户任务与“heredoc-file-write”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：`gdpval_bench/skills/heredoc-file-write/SKILL.md`
- **http-response-handling**（skill）：Handle websites requiring JavaScript by using curl with browser headers and validating file types. 激活提示：当用户任务与“http-response-handling”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：`gdpval_bench/skills/http-response-handling/SKILL.md`
- **incremental-excel-build**（skill）：Build complex Excel files through staged, verifiable steps with intermediate CSV outputs for debugging 激活提示：当用户任务与“incremental-excel-build”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：`gdpval_bench/skills/incremental-excel-build/SKILL.md`
- **irregular-excel-parsing**（skill）：Handle Excel files with irregular headers, merged cells, and unknown header row positions using pattern-matching and index-based extraction. 激活提示：当用户任务与“irregular-excel-parsing”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：`gdpval_bench/skills/irregular-excel-parsing/SKILL.md`
- **local-pdf-extraction**（skill）：Extract text from local PDFs using pdftotext or PyMuPDF via run shell 激活提示：当用户任务与“local-pdf-extraction”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：`gdpval_bench/skills/local-pdf-extraction/SKILL.md`
- **locate-missing-artifacts**（skill）：Recover generated files in sandboxed environments by searching from confirmed roots and copying to expected locations 激活提示：当用户任务与“locate-missing-artifacts”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：`gdpval_bench/skills/locate-missing-artifacts/SKILL.md`
- **medical-soap-note-creation**（skill）：Transform unstructured clinical encounters into comprehensive SOAP notes with ICD codes and care plans 激活提示：当用户任务与“medical-soap-note-creation”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：`gdpval_bench/skills/medical-soap-note-creation/SKILL.md`
- **multi-deliverable-tracking-1e842b**（skill）：Track completion status of all required deliverables and ensure ALL outputs are completed before stopping 激活提示：当用户任务与“multi-deliverable-tracking-1e842b”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：`gdpval_bench/skills/multi-deliverable-tracking-1e842b/SKILL.md`
- **multi-deliverable-tracking-2eb609**（skill）：Systematic tracking and completion of all required deliverables before task termination 激活提示：当用户任务与“multi-deliverable-tracking-2eb609”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：`gdpval_bench/skills/multi-deliverable-tracking-2eb609/SKILL.md`
- **multi-deliverable-tracking**（skill）：Track and complete all required deliverables before finishing a task 激活提示：当用户任务与“multi-deliverable-tracking”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：`gdpval_bench/skills/multi-deliverable-tracking/SKILL.md`
- **Skill**（skill）：The patch looks good and addresses the identified issues. Let me verify the skill file exists and then output the complete evolution. 激活提示：当用户任务与“Skill”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：`gdpval_bench/skills/multi-pass-video-editing/SKILL.md`
- **openpyxl-merged-cell-safe-write**（skill）：Safe pattern for writing values to merged cells in openpyxl spreadsheets 激活提示：当用户任务与“openpyxl-merged-cell-safe-write”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：`gdpval_bench/skills/openpyxl-merged-cell-safe-write/SKILL.md`
- **openpyxl-sandbox-workaround**（skill）：Use run shell with inline Python as a fallback when execute code sandbox fails for openpyxl/spreadsheet operations 激活提示：当用户任务与“openpyxl-sandbox-workaround”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：`gdpval_bench/skills/openpyxl-sandbox-workaround/SKILL.md`
- **pandoc-pdf-error-diagnosis**（skill）：Systematic diagnostic workflow for resolving pandoc PDF conversion errors 激活提示：当用户任务与“pandoc-pdf-error-diagnosis”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：`gdpval_bench/skills/pandoc-pdf-error-diagnosis/SKILL.md`
- **pandoc-pdf-error-recovery**（skill）：Systematic approach to diagnose and recover from pandoc PDF conversion unknown errors 激活提示：当用户任务与“pandoc-pdf-error-recovery”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：`gdpval_bench/skills/pandoc-pdf-error-recovery/SKILL.md`
- **pandoc-pdf-error-resolution**（skill）：Systematic debugging workflow for pandoc PDF conversion errors 激活提示：当用户任务与“pandoc-pdf-error-resolution”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：`gdpval_bench/skills/pandoc-pdf-error-resolution/SKILL.md`
- **pandoc-unicode-sanitize**（skill）：Pre-process Markdown to replace non-ASCII characters before pandoc PDF conversion 激活提示：当用户任务与“pandoc-unicode-sanitize”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：`gdpval_bench/skills/pandoc-unicode-sanitize/SKILL.md`
- **pandoc-unicode-workaround**（skill）：Handle LaTeX Unicode errors in pandoc PDF generation by normalizing special characters to ASCII 激活提示：当用户任务与“pandoc-unicode-workaround”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：`gdpval_bench/skills/pandoc-unicode-workaround/SKILL.md`
- **pdf-checklist-generation**（skill）：Generate structured PDF checklists and reports using Python libraries in sandbox 激活提示：当用户任务与“pdf-checklist-generation”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：`gdpval_bench/skills/pdf-checklist-generation/SKILL.md`
- **pdf-checklist-generator**（skill）：Generate structured PDF checklists and reports with tables, sections, and scoring criteria using Python libraries 激活提示：当用户任务与“pdf-checklist-generator”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：`gdpval_bench/skills/pdf-checklist-generator/SKILL.md`
- **unified-deliverable-workflow**（skill）：Generate spreadsheets, diagrams, and PDF reports with iteration budgeting and error recovery 激活提示：当用户任务与“unified-deliverable-workflow”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：`gdpval_bench/skills/pdf-checklist-workflow-enhanced-780c5d/SKILL.md`
- **integrated-report-bundle**（skill）：Generate spreadsheet, diagram, and PDF report as a cohesive deliverable set with iteration budgeting per component 激活提示：当用户任务与“integrated-report-bundle”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：`gdpval_bench/skills/pdf-checklist-workflow-enhanced-a53b26/SKILL.md`
- **unified-deliverables-flow**（skill）：Generate spreadsheets, diagrams, and PDF reports in a phased workflow with explicit iteration budgets and error recovery 激活提示：当用户任务与“unified-deliverables-flow”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：`gdpval_bench/skills/pdf-checklist-workflow-enhanced/SKILL.md`
- **pdf-checklist-workflow**（skill）：Generate structured PDF documents with tables, sections, and scoring using Python libraries in execute code sandbox 激活提示：当用户任务与“pdf-checklist-workflow”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：`gdpval_bench/skills/pdf-checklist-workflow/SKILL.md`
- **pdf-debug-unbuffered**（skill）：Debug Python PDF generation errors by using unbuffered output and stderr inspection to reveal actual tracebacks 激活提示：当用户任务与“pdf-debug-unbuffered”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：`gdpval_bench/skills/pdf-debug-unbuffered/SKILL.md`
- **pdf-extract-create-workflow**（skill）：Complete PDF lifecycle: download, extract, and generate structured documents with reportlab 激活提示：当用户任务与“pdf-extract-create-workflow”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：`gdpval_bench/skills/pdf-download-extract-fallback-enhanced-29b2f4/SKILL.md`
- **pdf-extract-progressive-tools**（skill）：Progressive tool-chain PDF extraction with explicit read file, run shell, and execute code sandbox sequencing 激活提示：当用户任务与“pdf-extract-progressive-tools”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：`gdpval_bench/skills/pdf-download-extract-fallback-enhanced-899f5b/SKILL.md`
- 其余 180 个条目见 `AI_CONTEXT_PACK.json`。

## 证据索引

- 共索引 180 条证据。

- **✨ OpenSpace: Make Your Agents: Smarter, Low-Cost, Self-Evolving ✨**（documentation）：✨ OpenSpace: Make Your Agents: Smarter, Low-Cost, Self-Evolving ✨ 证据：`README.md`
- **OpenSpace Frontend**（documentation）：A dashboard frontend for the OpenSpace, providing skill browsing, lineage visualization, and workflow inspection. 证据：`frontend/README.md`
- **GDPVal Benchmark**（documentation）：Benchmark for evaluating OpenSpace on GDPVal https://huggingface.co/datasets/openai/gdpval 220 occupational tasks across 44 occupations and 9 sectors . Measures token savings from skill accumulation by running each task twice: 证据：`gdpval_bench/README.md`
- **🖥️ My Daily Monitor**（documentation）：Your Entire Day on One Live Screen — with an AI Agent That Works for You 证据：`showcase/README.md`
- **🔧 Configuration Guide**（documentation）：!NOTE Create openspace/.env from .env.example ../../.env.example and set at least one LLM API key. 证据：`openspace/config/README.md`
- **Host Skills Integration Guide**（documentation）：This guide covers agent-specific setup for integrating OpenSpace. For installation and general concepts, see the main README ../../README.md -quick-start . 证据：`openspace/host_skills/README.md`
- **OpenSpace Local Server**（documentation）：The local server is a lightweight Flask service that runs on the host machine and exposes HTTP endpoints for shell execution and GUI automation. It is only needed in server mode — most users should use the default local mode instead. 证据：`openspace/local_server/README.md`
- **Skills**（documentation）：Place custom skills here. Each skill is a subdirectory containing a SKILL.md : 证据：`openspace/skills/README.md`
- **Package**（package_manifest）：{ "name": "openspace-frontend", "private": true, "version": "0.1.0", "type": "module", "scripts": { "dev": "vite --port 3789", "build": "tsc && vite build", "preview": "vite preview" }, "dependencies": { "axios": "^1.7.9", "i18next": "^26.0.3", "react": "^18.3.1", "react-dom": "^18.3.1", "react-force-graph-2d": "^1.25.4", "react-i18next": "^17.0.2", "react-router-dom": "^7.1.1" }, "devDependencies": { "@types/react": "^18.3.12", "@types/react-dom": "^18.3.1", "@vitejs/plugin-react": "^4.3.4", "autoprefixer": "^10.4.20", "postcss": "^8.4.49", "tailwindcss": "^3.4.17", "typescript": "~5.6.2", "vite": "^6.0.3" } } 证据：`frontend/package.json`
- **Package**（package_manifest）：{ "name": "openspace-whatsapp-bridge", "version": "1.0.0", "description": "WhatsApp bridge for OpenSpace using Baileys", "private": true, "type": "module", "scripts": { "start": "node bridge.js" }, "dependencies": { "@hapi/boom": "^10.0.1", "@whiskeysockets/baileys": "7.0.0-rc.9", "pino": "^9.0.0", "qrcode-terminal": "^0.12.0", "ws": "^8.18.0" } } 证据：`openspace/communication/bridges/whatsapp/package.json`
- **Package**（package_manifest）：{ "name": "my-daily-monitor", "private": true, "version": "0.2.0", "type": "module", "scripts": { "dev": "vite", "dev:api": "tsx watch server/index.ts", "dev:full": "concurrently \"npm run dev\" \"npm run dev:api\"", "build": "tsc && vite build", "preview": "vite preview" }, "devDependencies": { "concurrently": "^9.2.1", "tsx": "^4.21.0", "typescript": "^5.7.2", "vite": "^6.0.7" }, "dependencies": { "maplibre-gl": "^5.19.0" } } 证据：`showcase/my-daily-monitor/package.json`
- **Adaptive Stem Alignment Workflow**（skill_instruction）：This skill provides a resilient pattern for audio production that emphasizes incremental verification , fail-fast principles, and adaptive duration handling . Each major step produces verified outputs before proceeding, with explicit strategies for handling stems of different durations. 证据：`gdpval_bench/skills/audio-track-production-enhanced-enhanced-b8f537/SKILL.md`
- **Diagnostic Stem Audio Production Workflow**（skill_instruction）：Diagnostic Stem Audio Production Workflow 证据：`gdpval_bench/skills/audio-track-production-enhanced-enhanced-enhanced/SKILL.md`
- **Aligned Stem Audio Production Workflow**（skill_instruction）：Aligned Stem Audio Production Workflow 证据：`gdpval_bench/skills/audio-track-production-enhanced-enhanced/SKILL.md`
- **Incremental Audio Production Workflow**（skill_instruction）：Incremental Audio Production Workflow 证据：`gdpval_bench/skills/audio-track-production-enhanced/SKILL.md`
- **Audio Track Production Workflow**（skill_instruction）：This skill provides a reusable pattern for executing audio production tasks that require generating a master track and multiple stems, applying effects, and delivering verified outputs in an archive. 证据：`gdpval_bench/skills/audio-track-production/SKILL.md`
- **Cascade Failure Recovery**（skill_instruction）：When multiple data retrieval tools read webpage, search web fail in succession, immediately fall back to generating content from embedded knowledge rather than attempting repairs or abandoning the task. This preserves task context and ensures completion. 证据：`gdpval_bench/skills/cascade-fail-recovery/SKILL.md`
- **Code Execution Fallback Pattern**（skill_instruction）：Apply this pattern when you encounter repeated failures with execute code sandbox : - 2+ consecutive failures with opaque or unknown errors - Timeout errors that persist across retry attempts - Environment-related errors that don't resolve with code fixes 证据：`gdpval_bench/skills/code-exec-fallback-266cba/SKILL.md`
- **Code Execution Fallback**（skill_instruction）：Use this pattern when execute code sandbox fails repeatedly typically 2+ attempts due to environment limitations, timeouts, dependency issues, or sandbox restrictions. 证据：`gdpval_bench/skills/code-exec-fallback/SKILL.md`
- **Code Execution Fallback Workflow**（skill_instruction）：Use this skill when execute code sandbox fails repeatedly 2+ attempts with unknown, persistent, or unexplained errors. This fallback approach uses write file + run shell to save Python scripts to disk and execute them via command line, which has proven more reliable in certain failure scenarios. 证据：`gdpval_bench/skills/code-execution-fallback-e81068/SKILL.md`
- **Code Execution Fallback & Workspace Anchoring**（skill_instruction）：Code Execution Fallback & Workspace Anchoring 证据：`gdpval_bench/skills/code-execution-fallback/SKILL.md`
- **Comprehensive SOAP Note Creation**（skill_instruction）：This skill defines the workflow for creating structured medical documentation SOAP notes by writing all required sections to a file in one comprehensive operation. This ensures completeness, consistency, and efficiency in clinical documentation. 证据：`gdpval_bench/skills/create-soap-note-merged/SKILL.md`
- **Create SOAP Note**（skill_instruction）：This skill defines the workflow for creating structured medical documentation SOAP notes by writing comprehensive content directly to a file. It ensures all standard sections are included and properly formatted. 证据：`gdpval_bench/skills/create-soap-note/SKILL.md`
- **Debug Sandbox Execution Failures**（skill_instruction）：When execute code sandbox fails with unknown errors or incomplete output, use this debugging pattern to identify the root cause and recover incrementally. 证据：`gdpval_bench/skills/debug-sandbox-execution/SKILL.md`
- **Delegated File Edit Verification**（skill_instruction）：Use this workflow whenever another agent, subprocess, or delegated tool claims it successfully created or modified a file. The goal is to prevent false acceptance of incomplete, misplaced, or incorrect edits. 证据：`gdpval_bench/skills/delegated-file-edit-verification/SKILL.md`
- **Deliverable Completion Protocol**（skill_instruction）：This skill addresses a common misconception during document and file creation tasks: agents sometimes search for or attempt to use a submit work , finalize , or similar tool after creating the deliverable. No such step is required. Task completion is achieved when the file is successfully created with the required content. 证据：`gdpval_bench/skills/deliverable-completion/SKILL.md`
- **Detect Zero-Iteration Failure**（skill_instruction）：This skill provides a workflow for identifying and classifying agent failures that occur prior to any task execution. 证据：`gdpval_bench/skills/detect-zero-iteration-failure/SKILL.md`
- **Detect Zero-Iteration Failures**（skill_instruction）：This skill helps identify and properly categorize agent failures that occur before any task execution begins. These failures require system-level investigation rather than agent-level debugging. 证据：`gdpval_bench/skills/detect-zero-iteration-failures/SKILL.md`
- **Direct ReportLab PDF Generation via run shell**（skill_instruction）：Direct ReportLab PDF Generation via run shell 证据：`gdpval_bench/skills/direct-reportlab-pdf-generation/SKILL.md`
- **Resilient Document Generation Workflow Multi-Engine Fallback**（skill_instruction）：Resilient Document Generation Workflow Multi-Engine Fallback 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-2794b4/SKILL.md`
- **Document Generation with Diagnostic Workflow Unicode-Safe**（skill_instruction）：Document Generation with Diagnostic Workflow Unicode-Safe 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-3d3a9a/SKILL.md`
- **Resilient Document Generation Workflow Multi-Format + Unicode-Safe**（skill_instruction）：Resilient Document Generation Workflow Multi-Format + Unicode-Safe 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-618c9b/SKILL.md`
- **Resilient Document Generation Workflow**（skill_instruction）：Resilient Document Generation Workflow 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-96865f/SKILL.md`
- **Document Generation: Dual-Backend Workflow Unicode-Safe**（skill_instruction）：Document Generation: Dual-Backend Workflow Unicode-Safe 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-9f3b1f/SKILL.md`
- **Document Generation Fallback Workflow Unicode-Safe**（skill_instruction）：Document Generation Fallback Workflow Unicode-Safe 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced-c5d426/SKILL.md`
- **Resilient Document Generation Workflow**（skill_instruction）：Resilient Document Generation Workflow 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-enhanced/SKILL.md`
- **Resilient Document Generation Pipeline**（skill_instruction）：Resilient Document Generation Pipeline 证据：`gdpval_bench/skills/document-gen-fallback-enhanced-merged/SKILL.md`
- **Document Generation Fallback Workflow Unicode-Safe**（skill_instruction）：Document Generation Fallback Workflow Unicode-Safe 证据：`gdpval_bench/skills/document-gen-fallback-enhanced/SKILL.md`
- **Document Generation Fallback Workflow**（skill_instruction）：Document Generation Fallback Workflow 证据：`gdpval_bench/skills/document-gen-fallback/SKILL.md`
- **DOCX Read Fallback**（skill_instruction）：When read file or execute code sandbox fails to read .docx files, use run shell with python-docx as a reliable workaround. 证据：`gdpval_bench/skills/docx-read-fallback/SKILL.md`
- **DOCX Shell Extraction**（skill_instruction）：Use this pattern when you need to read or extract text from Microsoft Word .docx files in constrained environments where: - The python-docx library is not available - You cannot install additional Python packages - You need a quick, reliable shell-based solution 证据：`gdpval_bench/skills/docx-shell-extract/SKILL.md`
- **DOCX Dual-Method Text Extraction**（skill_instruction）：Extract text from Microsoft Word .docx files using either shell commands or Python's zipfile module, automatically selecting the most reliable method for your environment. 证据：`gdpval_bench/skills/docx-shell-parse-enhanced-4bba79/SKILL.md`
- **Resilient DOCX Text Extraction**（skill_instruction）：Extract text from Microsoft Word .docx files using a robust two-tier approach: shell-based extraction as the primary method, with Python zipfile fallback when shell commands fail or return no output. 证据：`gdpval_bench/skills/docx-shell-parse-enhanced/SKILL.md`
- **DOCX Shell Parsing Workaround**（skill_instruction）：When you need to read content from Microsoft Word .docx files but python-docx or similar libraries are unavailable, use this shell-based approach to extract text reliably. 证据：`gdpval_bench/skills/docx-shell-parse/SKILL.md`
- **DOCX Shell Workaround**（skill_instruction）：Use this skill when: - read file cannot extract content from .docx files - execute code sandbox encounters failures when creating or modifying docx files - You need a reliable fallback for docx file manipulation 证据：`gdpval_bench/skills/docx-shell-workaround/SKILL.md`
- **Excel File Creation Fallback**（skill_instruction）：When execute code sandbox fails for openpyxl or Excel operations, delegate to shell agent which can autonomously handle dependency and environment issues. 证据：`gdpval_bench/skills/excel-creation-fallback/SKILL.md`
- **Excel Debug-First Extraction Workflow**（skill_instruction）：Excel Debug-First Extraction Workflow 证据：`gdpval_bench/skills/excel-debug-extraction/SKILL.md`
- **Excel Creation via Shell Heredoc Fallback**（skill_instruction）：Excel Creation via Shell Heredoc Fallback 证据：`gdpval_bench/skills/excel-heredoc-fallback/SKILL.md`
- **Excel Heredoc Workaround**（skill_instruction）：When creating Excel files with openpyxl, execute code sandbox may fail due to environment issues. This skill provides a robust workaround: fall back to run shell with inline Python heredoc scripts to create properly formatted spreadsheets with styling. 证据：`gdpval_bench/skills/excel-heredoc-workaround/SKILL.md`
- **Excel Subtotal Audit**（skill_instruction）：Use this workflow after generating or modifying an Excel report to catch silent spreadsheet defects that simple file-exists or sheet-exists checks will miss. 证据：`gdpval_bench/skills/excel-subtotal-audit/SKILL.md`
- **Excel Unmerge Before Write**（skill_instruction）：When automating Excel file population with openpyxl, merged cells in templates can cause write failures. This skill provides the pattern to safely handle merged ranges before populating data. 证据：`gdpval_bench/skills/excel-unmerge-before-write/SKILL.md`
- **Execute Code Fallback Pattern**（skill_instruction）：When execute code sandbox fails due to e2b unavailability, immediately fall back to run shell with embedded Python code. This ensures task continuity when the sandbox service is down. 证据：`gdpval_bench/skills/execute-code-fallback/SKILL.md`
- **Fallback Code Execution Workflow**（skill_instruction）：This skill defines a robust workaround for executing code specifically Python when the primary execute code sandbox tool fails repeatedly with unknown or transient errors. Instead of continuing to retry the failing tool, the agent switches to a manual file-write and shell-execution pattern. 证据：`gdpval_bench/skills/fallback-code-execution/SKILL.md`
- **Fallback File Creation Workflow**（skill_instruction）：Use this skill when execute code sandbox or shell agent fails repeatedly for file creation tasks. This pattern provides a manual but reliable approach to create and verify files. 证据：`gdpval_bench/skills/fallback-file-creation/SKILL.md`
- **Fallback File Operations**（skill_instruction）：When execute code sandbox or shell agent fail with 'unknown error' on filesystem operations, fall back to run shell with explicit mkdir and file write commands. This hybrid approach is more reliable for creating complex project structures. 证据：`gdpval_bench/skills/fallback-file-ops/SKILL.md`
- **Fallback Python Execution Pattern**（skill_instruction）：Use this pattern when: - execute code sandbox returns unknown errors or fails repeatedly - shell agent cannot successfully execute Python code - You need to create files spreadsheets, documents, data files via Python - Direct delegated approaches prove unreliable in the current environment 证据：`gdpval_bench/skills/fallback-python-execution/SKILL.md`
- **Shell-Based Python Fallback**（skill_instruction）：Use this fallback pattern when: - execute code sandbox returns 'unknown error' - read file returns 'unknown error' for supported formats - You need to process documents, analyze data, or read files programmatically 证据：`gdpval_bench/skills/fallback-python-shell/SKILL.md`
- **Fallback Script Execution with write file + run shell**（skill_instruction）：Fallback Script Execution with write file + run shell 证据：`gdpval_bench/skills/fallback-script-execution/SKILL.md`
- **FFmpeg Encoder Availability Check**（skill_instruction）：Before writing any FFmpeg encoding script, always probe the system for available encoders. This prevents failures from missing or incompatible codec libraries especially libopenh264 which frequently has version mismatches . 证据：`gdpval_bench/skills/ffmpeg-encoder-check-4855c0/SKILL.md`
- **FFmpeg Encoder Availability Check**（skill_instruction）：Before writing video encoding scripts, always verify which H.264 encoders are available in your FFmpeg installation. This prevents failures from library version mismatches, particularly with libopenh264. 证据：`gdpval_bench/skills/ffmpeg-encoder-check/SKILL.md`
- 其余 120 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：`README.md`, `frontend/README.md`, `gdpval_bench/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：`README.md`, `frontend/README.md`, `gdpval_bench/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。

- **OpenSpace 概览与系统架构**：importance `high`
  - source_paths: README.md, README_CN.md, openspace/__init__.py, openspace/__main__.py, openspace/tool_layer.py
- **自演化技能引擎与云端社区**：importance `high`
  - source_paths: openspace/skill_engine/registry.py, openspace/skill_engine/evolver.py, openspace/skill_engine/analyzer.py, openspace/skill_engine/store.py, openspace/skill_engine/patch.py
- **MCP 集成、通信网关与前端 Dashboard**：importance `high`
  - source_paths: openspace/mcp_server.py, openspace/host_skills/README.md, openspace/host_skills/delegate-task/SKILL.md, openspace/host_skills/skill-discovery/SKILL.md, openspace/communication/gateway.py
- **多后端 Grounding 系统、安全沙箱、质量监控与运维排障**：importance `high`
  - source_paths: openspace/grounding/core/grounding_client.py, openspace/grounding/core/provider.py, openspace/grounding/core/session.py, openspace/grounding/core/search_tools.py, openspace/grounding/core/security/sandbox.py

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `228f8f78073dc4ed0e63fff01c19596c50115d40`
- inspected_files: `requirements.txt`, `pyproject.toml`, `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: 来源证据：Agents have OpenSpace MCP tools available but never call them — all skills show 0 score

- 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：Agents have OpenSpace MCP tools available but never call them — all skills show 0 score
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/HKUDS/OpenSpace/issues/84 | 来源讨论提到 macos 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 2: 来源证据：Why it is smoothly to execute the task in the openspace lonely, however, when it comes to the openclaw with openspace,…

- 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：Why it is smoothly to execute the task in the openspace lonely, however, when it comes to the openclaw with openspace, the mcp server always exceeds the limite…
- Why it matters: 可能阻塞安装或首次运行。
- Evidence: community_evidence:github | https://github.com/HKUDS/OpenSpace/issues/73 | 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：[Bug] OpenSpace interactive mode incompatible with launchd — EOFError loop, 100% CPU, 100GB+ logs

- Trigger: GitHub 社区证据显示该项目存在一个安装相关的待验证问题：[Bug] OpenSpace interactive mode incompatible with launchd — EOFError loop, 100% CPU, 100GB+ logs
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/HKUDS/OpenSpace/issues/87 | 来源讨论提到 python 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 4: 可能修改宿主 AI 配置

- Trigger: 项目面向 Claude/Cursor/Codex/Gemini/OpenCode 等宿主，或安装命令涉及用户配置目录。
- Host AI rule: 列出会写入的配置文件、目录和卸载/回滚步骤。
- Why it matters: 安装可能改变本机 AI 工具行为，用户需要知道写入位置和回滚方法。
- Evidence: capability.host_targets | github_repo:1190363679 | https://github.com/HKUDS/OpenSpace | host_targets=openclaw, mcp_host, claude_code, claude, cursor
- 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: 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:1190363679 | https://github.com/HKUDS/OpenSpace | 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 6: 运行可能依赖外部服务

- Trigger: 项目说明出现 external service/cloud/webhook/database 等运行依赖关键词。
- Host AI rule: 确认是否有离线 demo、mock 数据或可替代服务。
- Why it matters: 本地安装成功不等于能力可用，外部服务不可用会阻断体验。
- Evidence: packet_text.keyword_scan | github_repo:1190363679 | https://github.com/HKUDS/OpenSpace | matched external service / cloud / webhook / database keyword
- 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: 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:1190363679 | https://github.com/HKUDS/OpenSpace | 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 | github_repo:1190363679 | https://github.com/HKUDS/OpenSpace | 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 9: 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 | github_repo:1190363679 | https://github.com/HKUDS/OpenSpace | 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: 来源证据：Bump litellm pin to skip PYSEC-2026-2 (1.82.7+1.82.8) but allow 1.83.0+

- 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：Bump litellm pin to skip PYSEC-2026-2 (1.82.7+1.82.8) but allow 1.83.0+
- Why it matters: 可能影响授权、密钥配置或安全边界。
- Evidence: community_evidence:github | https://github.com/HKUDS/OpenSpace/issues/85 | 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。