Development

依各语言的最佳实践组织项目目录结构的指南。 Use when: 创建项目、重整结构、新增模块、配置构建流程、决定文件该放哪里。 Not for: 专门为 AI 导览而设计的结构——请用 /ai-friendly-architecture；原地重整既有代码——请用 /refactor。 Keywords: project, structure, directory, layout, gitignore, scaffold, file placement, utils, helpers, shared, 项目结构, 目录布局, 文件摆放.

设计一致的错误码，遵循 PREFIX_CATEGORY_NUMBER 格式。 使用时机：定義错误码、建立错误处理、设计 API。 关鍵字：error code, error handling, error format, API errors, 错误码, 错误处理。

从项目来源生成使用说明文档。 使用时机：「生成文档」、「创建速查表」、「使用指南」、「功能参考」、「列出所有功能」 关键字：documentation, usage, reference, cheatsheet, features, 功能文档, 速查表, 使用说明

引導提交前品质关卡和簽入工作流程。 使用时机：提交程序码、准备提交、品质关卡验证。 关鍵字：commit, checkin, pre-commit, quality gate, git add, 提交, 簽入, 品质关卡.

依照 Keep a Changelog 格式撰写与維護 CHANGELOG.md。 使用时机：建立变更日誌条目、准备發布、记录变更。 关鍵字：changelog, release notes, CHANGELOG.md, keep a changelog, 变更日誌, 發布说明。

[UDS] 行为驱动开发（BDD）的参考资料：Gherkin 的 Given-When-Then 格式与 Three Amigos 结构。 Use when: 编写或审查 .feature 场景、选定通用语言、针对行为进行探索式对话。 Not for: 执行 BDD 生命周期或 RED/GREEN 自动化——该部分已移至采用层（XSPEC-095）；把 .feature 文件转成 E2E 骨架——请用 /e2e。 Keywords: BDD, Gherkin, Given When Then, feature file, scenario, Three Amigos, 行为驱动开发, 场景, 特性文件, 通用语言.

[UDS] 验收测试驱动开发（ATDD）的参考资料：INVEST 准则、Gherkin 验收条件格式与 Three Amigos 结构。 Use when: 与 Product Owner 一起定义验收条件、进行规格工作坊、用 INVEST 检视用户故事。 Not for: 执行 ATDD 生命周期或强制 PO 签核关卡——该部分已移至采用层（XSPEC-095）；编写单元测试——请用 /tdd。 Keywords: ATDD, acceptance test, acceptance criteria, INVEST, specification workshop, Three Amigos, 验收测试驱动开发, 验收条件, 规格工作坊.

创建和维护 AI 指令文件（CLAUDE.md、.cursorrules 等）并采用适当结构。 使用时机：创建 AI 指令文件、区分通用与项目特定规则、配置 AI 工具。 关键字：CLAUDE.md, cursorrules, windsurfrules, clinerules, AI instructions, system prompt, 指令文件, AI 设定.

设计 AI 友善架构，包含明确模式、分层文档和语义边界。 使用时机：为 AI 协作规划项目结构、优化代码库以利 AI 分析、设置 AI 上下文。 关键字：architecture, AI-friendly, context, modules, documentation layers, .ai-context.yaml, 架构, AI 友善, 上下文.

**Priority**: P1 **Status**: Draft **Last Updated**: 2026-01-28 **Feature ID**: SKILL-VIBE-001 **Dependencies**: [SPEC-VIBE-01 Vibe Coding Integration] ---

Build per-H3 writer context packs (NO PROSE): merge briefs + evidence packs + anchor facts + allowed citations into a single deterministic JSONL, so drafting is less hollow and less brittle.

Initialize a new workspace by copying the standard artifact template (STATUS.md, CHECKPOINTS.md, UNITS.csv, DECISIONS.md + folders). **Trigger**: workspace init, initialize workspace, workspace template, 初始化 workspace. **Use when**: 启动任何 pipeline run（必须先有 workspace 工件与目录骨架）。 **Skip if**: workspace 已初始化且不希望覆盖既有文件（除非显式 `--overwrite`）。 **Network**: none. **Guardrail**: 不要修改 `.codex/skills/workspace-init/assets/` 模板；默认不覆盖已有文件。

Instantiate or update a workspace `UNITS.csv` from a selected pipeline and units template (deps/checkpoints/acceptance).

Execute exactly one eligible Unit in an existing research Workspace; use for stepwise or manual semantic execution when status, Attempt, Artifact, Manifest, checkpoint, and acceptance evidence must remain synchronized.

Generate lightweight section/subsection transitions (NO NEW FACTS) to prevent “island” subsections; outputs a transition map that merging/writing can weave in.

Normalize terminology across a draft (canonical terms + synonym policy) without changing citations or meaning.

Build a 2+ level taxonomy (`outline/taxonomy.yml`) from a core paper set and scope constraints, with short descriptions per node.

Fill `outline/tables_index.md` from `outline/table_schema.md` + evidence packs (NO PROSE in cells; citation-backed rows).

Identify survey/review papers in a retrieved set and extract taxonomy seeds into `outline/taxonomy.yml` (topics/subtopics/terminology).

Polish a single H3 unit file under `sections/` into survey-grade prose (de-template + contrast/eval/limitation), without changing citation keys.

Build per-subsection writing briefs (NO PROSE) so later drafting is driven by evidence and checkable comparison axes (not outline placeholders).

De-slot and harmonize paper voice across `sections/*.md` without changing meaning or citation keys.

Map papers from the core set to each outline subsection and write `outline/mapping.tsv` with coverage tracking.

Logic coherence pass for per-H3 section files: enforce a clear paragraph-1 thesis and surface paragraph-island risks (connector stats are diagnostic, not a quota) before merging.