|
|
@@ -1,434 +0,0 @@
|
|
|
-developer_guidelines:
|
|
|
- metadata:
|
|
|
- version: "1.2"
|
|
|
- last_updated: "2025-10-24"
|
|
|
- purpose: "统一开发与自动化行为规范；在文件生成、推送流程与工程决策中落实可执行的核心哲学与强约束规则"
|
|
|
-
|
|
|
- principles:
|
|
|
- interface_handling:
|
|
|
- id: "P1"
|
|
|
- title: "接口处理"
|
|
|
- rules:
|
|
|
- - "所有接口调用或实现前，必须查阅官方或内部文档"
|
|
|
- - "禁止在未查阅文档的情况下猜测接口、参数或返回值"
|
|
|
- - "接口行为必须通过权威来源确认（文档、代码、接口说明）"
|
|
|
- execution_confirmation:
|
|
|
- id: "P2"
|
|
|
- title: "执行确认"
|
|
|
- rules:
|
|
|
- - "在执行任何任务前，必须明确输入、输出、边界与预期结果"
|
|
|
- - "若存在任何不确定项，必须在执行前寻求确认"
|
|
|
- - "禁止在边界不清或需求模糊的情况下开始实现"
|
|
|
- business_understanding:
|
|
|
- id: "P3"
|
|
|
- title: "业务理解"
|
|
|
- rules:
|
|
|
- - "所有业务逻辑必须来源于明确的需求说明或人工确认"
|
|
|
- - "禁止基于个人假设或推测实现业务逻辑"
|
|
|
- - "需求确认过程必须留痕，以供追溯"
|
|
|
- code_reuse:
|
|
|
- id: "P4"
|
|
|
- title: "代码复用"
|
|
|
- rules:
|
|
|
- - "在创建新模块、接口或函数前，必须检查现有可复用实现"
|
|
|
- - "若现有实现可满足需求，必须优先复用"
|
|
|
- - "禁止在已有功能满足需求时重复开发"
|
|
|
- quality_assurance:
|
|
|
- id: "P5"
|
|
|
- title: "质量保证"
|
|
|
- rules:
|
|
|
- - "提交代码前，必须具备可执行的测试用例"
|
|
|
- - "所有关键逻辑必须通过单元测试或集成测试验证"
|
|
|
- - "禁止在未通过测试的情况下提交或上线代码"
|
|
|
- architecture_compliance:
|
|
|
- id: "P6"
|
|
|
- title: "架构规范"
|
|
|
- rules:
|
|
|
- - "必须遵循现行架构规范与约束"
|
|
|
- - "禁止修改架构层或跨层调用未授权模块"
|
|
|
- - "任何架构变更需经负责人或架构评审批准"
|
|
|
- honest_communication:
|
|
|
- id: "P7"
|
|
|
- title: "诚信沟通"
|
|
|
- rules:
|
|
|
- - "在理解不充分或信息不完整时，必须主动说明"
|
|
|
- - "禁止假装理解、隐瞒不确定性或未经确认即执行"
|
|
|
- - "所有关键沟通必须有记录"
|
|
|
- code_modification:
|
|
|
- id: "P8"
|
|
|
- title: "代码修改"
|
|
|
- rules:
|
|
|
- - "在修改代码前，必须分析依赖与影响范围"
|
|
|
- - "必须保留回退路径并验证改动安全性"
|
|
|
- - "禁止未经评估直接修改核心逻辑或公共模块"
|
|
|
-
|
|
|
-automation_rules:
|
|
|
- file_header_generation:
|
|
|
- description: "所有新生成的代码或文档文件都必须包含标准文件头说明；根据各自语法生成/嵌入注释或采用替代策略。"
|
|
|
- rule:
|
|
|
- - "支持注释语法的文件：按 language_comment_styles 渲染 inline_file_header_template 并插入到文件顶部。"
|
|
|
- - "不支持注释语法的文件（如 json/csv/parquet/xlsx/pdf/png/jpg 等）：默认生成旁挂元数据文件 `<filename>.meta.md`，写入同样内容；如明确允许 JSONC/前置 Front-Matter，则按 `non_comment_formats.strategy` 执行。"
|
|
|
- - "禁止跳过或忽略文件头生成步骤；CI/钩子需校验头注释或旁挂元数据是否存在且时间戳已更新。"
|
|
|
- - "文件头中的占位符（如 {自动生成时间}）必须在生成时实际替换为具体值。"
|
|
|
- language_detection:
|
|
|
- strategy: "优先依据文件扩展名识别语言；若无法识别，则尝试基于内容启发式判定；仍不确定时回退为 'sidecar_meta' 策略。"
|
|
|
- fallback: "sidecar_meta"
|
|
|
- language_comment_styles:
|
|
|
- # 单行注释类（逐行加前缀）
|
|
|
- - exts: [".py"] # Python
|
|
|
- style: "line"
|
|
|
- line_prefix: "# "
|
|
|
- - exts: [".sh", ".bash", ".zsh"] # Shell
|
|
|
- style: "line"
|
|
|
- line_prefix: "# "
|
|
|
- - exts: [".rb"] # Ruby
|
|
|
- style: "line"
|
|
|
- line_prefix: "# "
|
|
|
- - exts: [".rs"] # Rust
|
|
|
- style: "line"
|
|
|
- line_prefix: "// "
|
|
|
- - exts: [".go"] # Go
|
|
|
- style: "line"
|
|
|
- line_prefix: "// "
|
|
|
- - exts: [".ts", ".tsx", ".js", ".jsx"] # TS/JS
|
|
|
- style: "block"
|
|
|
- block_start: "/*"
|
|
|
- line_prefix: " * "
|
|
|
- block_end: "*/"
|
|
|
- - exts: [".java", ".kt", ".scala", ".cs"] # JVM/C#
|
|
|
- style: "block"
|
|
|
- block_start: "/*"
|
|
|
- line_prefix: " * "
|
|
|
- block_end: "*/"
|
|
|
- - exts: [".c", ".h", ".cpp", ".hpp", ".cc"] # C/C++
|
|
|
- style: "block"
|
|
|
- block_start: "/*"
|
|
|
- line_prefix: " * "
|
|
|
- block_end: "*/"
|
|
|
- - exts: [".css"] # CSS
|
|
|
- style: "block"
|
|
|
- block_start: "/*"
|
|
|
- line_prefix: " * "
|
|
|
- block_end: "*/"
|
|
|
- - exts: [".sql"] # SQL
|
|
|
- style: "line"
|
|
|
- line_prefix: "-- "
|
|
|
- - exts: [".yml", ".yaml", ".toml", ".ini", ".cfg"] # 配置类
|
|
|
- style: "line"
|
|
|
- line_prefix: "# "
|
|
|
- - exts: [".md"] # Markdown
|
|
|
- style: "block"
|
|
|
- block_start: "<!--"
|
|
|
- line_prefix: " "
|
|
|
- block_end: "-->"
|
|
|
- - exts: [".html", ".xml"] # HTML/XML
|
|
|
- style: "block"
|
|
|
- block_start: "<!--"
|
|
|
- line_prefix: " "
|
|
|
- block_end: "-->"
|
|
|
- non_comment_formats:
|
|
|
- formats: [".json", ".csv", ".parquet", ".xlsx", ".pdf", ".png", ".jpg", ".jpeg", ".gif"]
|
|
|
- strategy:
|
|
|
- json:
|
|
|
- preferred: "jsonc_if_allowed" # 若项目明确接受 JSONC/配置文件可带注释，则使用 /* ... */ 样式写 JSONC
|
|
|
- otherwise: "sidecar_meta" # 否则写 `<filename>.meta.md`
|
|
|
- csv: "sidecar_meta"
|
|
|
- parquet: "sidecar_meta"
|
|
|
- xlsx: "sidecar_meta"
|
|
|
- binary_default: "sidecar_meta" # 其余二进制/不可注释格式
|
|
|
- inline_file_header_template: |
|
|
|
- ############################################################
|
|
|
- # 📘 文件说明：
|
|
|
- # 本文件实现的功能：简要描述该代码文件的核心功能、作用和主要模块。
|
|
|
- #
|
|
|
- # 📋 程序整体伪代码（中文）：
|
|
|
- # 1. 初始化主要依赖与变量；
|
|
|
- # 2. 加载输入数据或接收外部请求；
|
|
|
- # 3. 执行主要逻辑步骤（如计算、处理、训练、渲染等）；
|
|
|
- # 4. 输出或返回结果；
|
|
|
- # 5. 异常处理与资源释放；
|
|
|
- #
|
|
|
- # 🔄 程序流程图（逻辑流）：
|
|
|
- # ┌──────────┐
|
|
|
- # │ 输入数据 │
|
|
|
- # └─────┬────┘
|
|
|
- # ↓
|
|
|
- # ┌────────────┐
|
|
|
- # │ 核心处理逻辑 │
|
|
|
- # └─────┬──────┘
|
|
|
- # ↓
|
|
|
- # ┌──────────┐
|
|
|
- # │ 输出结果 │
|
|
|
- # └──────────┘
|
|
|
- #
|
|
|
- # 📊 数据管道说明：
|
|
|
- # 数据流向：输入源 → 数据清洗/转换 → 核心算法模块 → 输出目标（文件 / 接口 / 终端）
|
|
|
- #
|
|
|
- # 🧩 文件结构：
|
|
|
- # - 模块1：xxx 功能；
|
|
|
- # - 模块2：xxx 功能；
|
|
|
- # - 模块3：xxx 功能；
|
|
|
- #
|
|
|
- # 🕒 创建时间：{自动生成时间}
|
|
|
- # 👤 作者/责任人：{author}
|
|
|
- # 🔖 版本：{version}
|
|
|
- ############################################################
|
|
|
-
|
|
|
- file_creation_compliance:
|
|
|
- description: "所有新文件的创建位置与结构必须符合内部文件生成规范"
|
|
|
- rule:
|
|
|
- - "文件生成逻辑必须遵循 inline_file_gen_spec 中的规定（已内联）"
|
|
|
- - "文件输出路径、模块层级、命名约定等均应匹配规范定义"
|
|
|
- - "不得在规范之外的位置生成文件"
|
|
|
- - "绝对禁止在项目根目录生成任何非文档规范可以出现的文件"
|
|
|
- inline_file_gen_spec:
|
|
|
- goal: "统一 AI 生成内容（文档、代码、测试文件等）的结构与路径，避免污染根目录或出现混乱命名。"
|
|
|
- project_structure: |
|
|
|
- project_root/
|
|
|
- │
|
|
|
- ├── docs/ # 📘 文档区
|
|
|
- │ ├── spec/ # 规范化文档（AI生成放这里）
|
|
|
- │ ├── design/ # 设计文档、接口文档
|
|
|
- │ └── readme.md
|
|
|
- │
|
|
|
- ├── src/ # 💻 源代码区
|
|
|
- │ ├── core/ # 核心逻辑
|
|
|
- │ ├── api/ # 接口层
|
|
|
- │ ├── utils/ # 工具函数
|
|
|
- │ └── main.py (或 index.js)
|
|
|
- │
|
|
|
- ├── tests/ # 🧪 单元测试
|
|
|
- │ ├── test_core.py
|
|
|
- │ └── test_api.py
|
|
|
- │
|
|
|
- ├── configs/ # ⚙️ 配置文件
|
|
|
- │ ├── settings.yaml
|
|
|
- │ └── logging.conf
|
|
|
- │
|
|
|
- ├── scripts/ # 🛠️ 自动化脚本、AI集成脚本
|
|
|
- │ └── generate_docs.py # （AI自动生成文档脚本）
|
|
|
- │
|
|
|
- ├── data/ # 📂 数据集、样例输入输出
|
|
|
- │
|
|
|
- ├── output/ # 临时生成文件、导出文件
|
|
|
- │
|
|
|
- ├── CLAUDE.md # CLAUDE记忆文件
|
|
|
- │
|
|
|
- ├── .gitignore
|
|
|
- ├── requirements.txt / package.json
|
|
|
- └── README.md
|
|
|
- generation_rules:
|
|
|
- - file_type: "Python 源代码"
|
|
|
- path: "/src"
|
|
|
- naming: "模块名小写，下划线分隔"
|
|
|
- notes: "遵守 PEP8"
|
|
|
- - file_type: "测试代码"
|
|
|
- path: "/tests"
|
|
|
- naming: "test_模块名.py"
|
|
|
- notes: "使用 pytest 格式"
|
|
|
- - file_type: "文档（Markdown）"
|
|
|
- path: "/docs"
|
|
|
- naming: "模块名_说明.md"
|
|
|
- notes: "UTF-8 编码"
|
|
|
- - file_type: "临时输出或压缩包"
|
|
|
- path: "/output"
|
|
|
- naming: "自动生成时间戳后缀"
|
|
|
- notes: "可被自动清理"
|
|
|
- coding_standards:
|
|
|
- style:
|
|
|
- - "严格遵守 PEP8"
|
|
|
- - "函数名用小写加下划线；类名大驼峰；常量全大写"
|
|
|
- docstrings:
|
|
|
- - "每个模块包含模块级 docstring"
|
|
|
- - "函数注明参数与返回类型（Google 或 NumPy 风格）"
|
|
|
- imports_order:
|
|
|
- - "标准库"
|
|
|
- - "第三方库"
|
|
|
- - "项目内模块"
|
|
|
- ai_generation_conventions:
|
|
|
- - "不得在根目录创建文件"
|
|
|
- - "所有新文件必须放入正确的分类文件夹"
|
|
|
- - "文件名应具有可读性与语义性"
|
|
|
- - defaults:
|
|
|
- code: "/src"
|
|
|
- tests: "/tests"
|
|
|
- docs: "/docs"
|
|
|
- temp: "/output"
|
|
|
- repository_push_rules:
|
|
|
- description: "所有推送操作必须符合远程仓库推送规范"
|
|
|
- rule:
|
|
|
- - "每次推送至远程仓库前，必须遵循 inline_repo_push_spec 的流程（已内联）"
|
|
|
- - "推送操作必须遵循其中定义的 GitHub 环境变量与流程说明"
|
|
|
- - "禁止绕过该流程进行直接推送"
|
|
|
- inline_repo_push_spec:
|
|
|
- github_env:
|
|
|
- GITHUB_ID: "https://github.com/xxx"
|
|
|
- GITHUB_KEYS: "ghp_xxx"
|
|
|
- core_principles:
|
|
|
- - "自动化"
|
|
|
- - "私有化"
|
|
|
- - "时机恰当"
|
|
|
- naming_rule: "改动的上传命名和介绍要以改动了什么，处于什么阶段和环境"
|
|
|
- triggers:
|
|
|
- on_completion:
|
|
|
- - "代码修改完成并验证"
|
|
|
- - "功能实现完成"
|
|
|
- - "错误修复完成"
|
|
|
- pre_risky_change:
|
|
|
- - "大规模代码重构前"
|
|
|
- - "删除核心功能或文件前"
|
|
|
- - "实验性高风险功能前"
|
|
|
- required_actions:
|
|
|
- - "优先提交所有变更（commit）并推送（push）到远程私有仓库"
|
|
|
- safety_policies:
|
|
|
- - "仅推送到私有仓库"
|
|
|
- - "新仓库必须设为 Private"
|
|
|
- - "禁止任何破坏仓库的行为与命令"
|
|
|
-
|
|
|
- core_philosophy:
|
|
|
- good_taste:
|
|
|
- id: "CP1"
|
|
|
- title: "好品味（消除特殊情况）"
|
|
|
- mandates:
|
|
|
- - "通过更通用建模消除特殊情况；能重构就不加分支"
|
|
|
- - "等价逻辑选择更简洁实现"
|
|
|
- - "评审审视是否有更通用模型"
|
|
|
- notes:
|
|
|
- - "例：链表删除逻辑改为无条件统一路径"
|
|
|
- never_break_userspace:
|
|
|
- id: "CP2"
|
|
|
- title: "不破坏用户空间（向后兼容）"
|
|
|
- mandates:
|
|
|
- - "导致现有程序崩溃或行为改变的变更默认是缺陷"
|
|
|
- - "接口变更需提供兼容层或迁移路径"
|
|
|
- - "合并前完成兼容性评估与回归"
|
|
|
- pragmatism:
|
|
|
- id: "CP3"
|
|
|
- title: "实用主义（问题导向）"
|
|
|
- mandates:
|
|
|
- - "优先解决真实问题，避免过度设计"
|
|
|
- - "性能/可维护性/时效做量化权衡并记录"
|
|
|
- - "拒绝为“理论完美”显著提升复杂度"
|
|
|
- simplicity_doctrine:
|
|
|
- id: "CP4"
|
|
|
- title: "简洁执念（控制复杂度）"
|
|
|
- mandates:
|
|
|
- - "函数单一职责；圈复杂度≤10"
|
|
|
- - "最大嵌套层级≤3，超出需重构或拆分"
|
|
|
- - "接口与命名精炼、语义明确"
|
|
|
- - "新增复杂度需设计说明与测试覆盖"
|
|
|
- cognitive_protocol:
|
|
|
- id: "CP5"
|
|
|
- title: "深度思考协议（UltraThink）"
|
|
|
- mandates:
|
|
|
- - "重要变更前执行 UltraThink 预检：问题重述→约束与目标→边界与反例→更简模型→风险与回退"
|
|
|
- - "预检结论记录在变更描述或提交信息"
|
|
|
- - "鼓励采用 SOTA，前提是不破坏 CP2 与 P6"
|
|
|
- excellence_bar:
|
|
|
- id: "CP6"
|
|
|
- title: "STOA 追求（State-of-the-Art）"
|
|
|
- mandates:
|
|
|
- - "关键路径对标 SOTA 并记录差距与收益"
|
|
|
- - "引入前沿方法需收益评估、替代对比、回退方案"
|
|
|
- - "禁止为新颖性牺牲稳定性与可维护性"
|
|
|
- Extremely_deep_thinking:
|
|
|
- id: "CP7"
|
|
|
- title: "极致深度思考（Extremely_deep_thinking:）"
|
|
|
- mandates:
|
|
|
- - "每次操作文件前进行深度思考，追求卓越产出"
|
|
|
- - "ultrathink ultrathink ultrathink ultrathink"
|
|
|
- - "STOA(state-of-the-art) 重复强调"
|
|
|
-
|
|
|
- usage_scope:
|
|
|
- applies_to:
|
|
|
- - "API接口开发与调用"
|
|
|
- - "业务逻辑实现"
|
|
|
- - "代码重构与优化"
|
|
|
- - "架构设计与调整"
|
|
|
- - "自动文件生成"
|
|
|
- - "Git推送与持续集成"
|
|
|
-
|
|
|
- pre_execution_checklist:
|
|
|
- - "已查阅相关文档并确认接口规范（P1）"
|
|
|
- - "已明确任务边界与输出预期（P2）"
|
|
|
- - "已核对可复用模块或代码（P4）"
|
|
|
- - "已准备测试方案或用例并通过关键用例（P5）"
|
|
|
- - "已确认符合架构规范与审批要求（P6）"
|
|
|
- - "已根据自动化规则加载并遵循三份规范（已内联版）"
|
|
|
- - "已完成 UltraThink 预检并记录结论（CP5）"
|
|
|
- - "已执行兼容性影响评估：不得破坏用户空间（CP2）"
|
|
|
- - "最大嵌套层级 ≤ 3，函数单一职责且复杂度受控（CP4）"
|
|
|
-
|
|
|
-prohibited_git_operations:
|
|
|
- history_rewriting:
|
|
|
- - command: "git push --force / -f"
|
|
|
- reason: "强制推送覆盖远程历史，抹除他人提交"
|
|
|
- alternative: "正常 git push；冲突用 merge 或 revert"
|
|
|
- - command: "git push origin main --force"
|
|
|
- reason: "重写主分支历史，风险极高"
|
|
|
- alternative: "git revert 针对性回滚"
|
|
|
- - command: "git commit --amend（已推送提交）"
|
|
|
- reason: "修改已公开历史破坏一致性"
|
|
|
- alternative: "新增提交补充说明"
|
|
|
- - command: "git rebase（公共分支）"
|
|
|
- reason: "改写历史导致协作混乱"
|
|
|
- alternative: "git merge"
|
|
|
- branch_structure:
|
|
|
- - command: "git branch -D main"
|
|
|
- reason: "强制删除主分支"
|
|
|
- alternative: "禁止删除主分支"
|
|
|
- - command: "git push origin --delete main"
|
|
|
- reason: "删除远程主分支导致仓库不可用"
|
|
|
- alternative: "禁止操作"
|
|
|
- - command: "git reset --hard HEAD~n"
|
|
|
- reason: "回滚并丢弃修改"
|
|
|
- alternative: "逐步使用 git revert"
|
|
|
- - command: "git reflog expire ... + git gc --prune=now --aggressive"
|
|
|
- reason: "彻底清理历史，几乎不可恢复"
|
|
|
- alternative: "禁止对 .git 进行破坏性清理"
|
|
|
- repo_polution_damage:
|
|
|
- - behavior: "删除 .git"
|
|
|
- reason: "失去版本追踪"
|
|
|
- alternative: "禁止删除；需要新项目请新路径初始化"
|
|
|
- - behavior: "将远程改为公共仓库"
|
|
|
- reason: "私有代码泄露风险"
|
|
|
- alternative: "仅使用私有仓库 URL"
|
|
|
- - behavior: "git filter-branch（不熟悉）"
|
|
|
- reason: "改写历史易误删敏感信息"
|
|
|
- alternative: "禁用；由管理员执行必要清理"
|
|
|
- - behavior: "提交 .env/API key/密钥"
|
|
|
- reason: "敏感信息泄露"
|
|
|
- alternative: "使用 .gitignore 与安全变量注入"
|
|
|
- external_risks:
|
|
|
- - behavior: "未验证脚本/CI 执行 git push"
|
|
|
- reason: "可能推送未审核代码或错误配置"
|
|
|
- alternative: "仅允许内部安全脚本执行"
|
|
|
- - behavior: "公共终端/云服务器保存 GITHUB_KEYS"
|
|
|
- reason: "极高泄露风险"
|
|
|
- alternative: "仅存放于安全环境变量中"
|
|
|
- - behavior: "root 强制清除 .git"
|
|
|
- reason: "版本丢失与协作混乱"
|
|
|
- alternative: "禁止；必要时新仓库备份迁移"
|
|
|
- collaboration_issues:
|
|
|
- - behavior: "直接在主分支提交"
|
|
|
- reason: "破坏审查机制，难以追踪来源"
|
|
|
- alternative: "feature 分支 → PR → Merge"
|
|
|
- - behavior: "未同步远程更新前直接推送"
|
|
|
- reason: "易造成冲突与历史分歧"
|
|
|
- alternative: "每次提交前先 git pull"
|
|
|
- - behavior: "将本地测试代码推到主分支"
|
|
|
- reason: "污染生产"
|
|
|
- alternative: "测试代码仅在 test/ 分支"
|
|
|
-
|
|
|
-git_safe_practices:
|
|
|
- - "在 git pull 前确认冲突风险（必要时 --rebase，但需评估）"
|
|
|
- - "历史修改、清理、合并在单独分支并经管理员审核"
|
|
|
- - "高风险操作前强制自动备份"
|
|
|
-
|
|
|
-appendices:
|
|
|
- ai_generation_spec_markdown: |
|
|
|
- # 🧠 AI 文件与代码生成规范记忆文档（原始说明保留）
|
|
|
- （已上方结构化到 inline_file_gen_spec，这里保留原始 Markdown 作参考）
|
|
|
-
|
|
|
- file_header_template_text: |
|
|
|
- （已上方结构化到 automation_rules.file_header_generation.inline_file_header_spec）
|
tukuaiai