数据库迁移
- 作者仓库星标 0
- 作者更新于 实时读取
- 作者仓库 skills-registry
- 领域
- 文档
- 兼容 Agent
-
- Claude Code
- Cursor
- Cline
- Codex
- Windsurf
- Gemini CLI
- +20
- 信任分
- 88 / 100 · 社区维护
- 作者 / 版本 / 许可
- @tomevault-io · 未声明 license
- Token 消耗评级
- 低消耗
- 接入复杂程度
- 需简单配置
- 是否需要外部 API Key
- 不需要
- 兼容的系统
- 未声明(默认跨平台)
- 底层运行要求
- 无特殊要求
- 文件与系统权限
-
- 只读
- 允许写入 / 修改
- Shell 执行
- 网络行为
- 仅限本地
- 安装命令数
- 26 条
档案由构建时根据 SKILL.md 与安装命令自动衍生,可能与作者实际意图存在差异。
需要注意: 未限定 allowed-tools,默认拥有全部工具权限。
---
name: agdmigrate-agents-md
description: Migrate existing agent documentation to the standardized AGENTS.md format. Use when the user wan…
category: 文档
runtime: 无特殊运行时
---
# agdmigrate-agents-md 输出预览
## PART A: 任务判断
- 适用问题:PRD、RFC、README、项目说明或知识库整理。
- 输入要求:目标材料、限制条件、期望输出和验收方式。
- 证据边界:围绕“Load Context / Migration Sources / Process”读取原文规则,不把推断写成作者承诺。
## PART B: 执行结果
- **01** 任务判断:确认你的需求是否属于PRD、RFC、README、项目说明或知识库整理,并标出输入、限制和预期结果。
- **02** 执行计划:优先按“Load Context / Migration Sources / Process”拆成步骤,说明每一步会读取什么、修改什么、产出什么。
- **03** 交付结果:给出可复制的命令、文件改动、检查清单或内容草稿,并说明如何继续迭代。
- **04** 风险边界:结合 读取文件、写入/修改文件、执行终端命令、主要在本地完成、通常不需要额外 API Key 给出执行前确认项。
## Running Rules
- 读取文件、写入/修改文件、执行终端命令;主要在本地完成;通常不需要额外 API Key。
- 先小样例验证,再放大到真实任务。
- 交付时同时给结果、检查口径和下一步迭代建议。 原文没有稳定的斜杠命令要求。安装验证后通常全局生效,直接在对话里点名这个 Skill 并描述任务即可。
告诉 Agent 目标文件或材料、期望结果、不可改范围、是否允许联网或执行命令。本 Skill 的权限画像是:读取文件、写入/修改文件、执行终端命令。
先用一个小任务确认它会围绕“Load Context / Migration Sources / Process”工作;涉及文件或命令时,先看 diff、日志、预览或测试结果。
检查最终产物是否包含明确结果、必要证据和下一步动作;如果输出泛泛而谈,就补充输入、边界和验收标准后重跑。
---
name: agdmigrate-agents-md
description: Migrate existing agent documentation to the standardized AGENTS.md format. Use when the user wan…
category: 文档
source: tomevault-io/skills-registry
---
# agdmigrate-agents-md
## 什么时候使用
- 把项目文档方向的常用动作沉淀成 Agent 可调用的技能 适合处理README、PRD、RFC、教程和知识库文档,核心价值是把输入、判断、执行、验证和交付边界固定下来,避免 Agent 泛泛回答。 把任务拆成可执行、可检查、可继续迭代的…
- 面向PRD、RFC、README、项目说明或知识库整理,优先处理能明确输入、步骤和验收标准的工作。
## 需要提供什么
- 目标材料、目录范围、期望结果和不可改动内容。
- 是否允许联网、执行命令、读写文件或调用外部服务。
## 执行规则
- 围绕「Load Context / Migration Sources / Process」组织步骤,不把推断写成作者事实。
- 读取文件、写入/修改文件、执行终端命令;主要在本地完成;通常不需要额外 API Key。
- 先跑小样例,确认结果可检查后再扩大任务范围。
## 输出要求
- 给出最终产物、关键证据、验证方式和下一步动作。
- 信息不足时标记 unknown,不编造命令、平台或依赖。 作者原文负责流程事实;仓库文件负责来源和命令;流狐只补充适用场景、限制和质量判断。
skill "agdmigrate-agents-md" {
输入层 -> 用户目标 + 目标文件 + 禁止范围 + 验收标准
上下文层 -> Load Context / Migration Sources / Process
规则层 -> SKILL.md 触发条件 / 执行顺序 / 输出格式
运行层 -> 无特殊运行时 | 读取文件、写入/修改文件、执行终端命令 | 主要在本地完成
安全层 -> 通常不需要额外 API Key + 小任务验证 + diff / 日志复核
输出层 -> 可复制结果 + 检查清单 + 下一步迭代
} Migrate existing agent documentation files to the standardized AGENTS.md format. Default to creating a compact single file; ask about references folder only for complex projects.
Load Context
Load the agents-md skill for format guidance:
Use Skill tool with skill="agd:agents-md"
Provides: sections, structure, symlink patterns, best practices
Migration Sources
Detect and migrate from common agent documentation files:
| Source File | Detection | Migration Strategy |
|---|---|---|
AGENT.md |
Single-file predecessor | Convert to AGENTS.md, ensure all sections covered |
.cursorrules |
Cursor IDE rules | Extract rules, convert to AGENTS.md sections |
.windsurfrules |
Windsurf IDE rules | Extract rules, convert to AGENTS.md sections |
CLAUDE.md |
Claude Code instructions | Merge with AGENTS.md or convert |
contributing.md |
Developer guidelines | Extract relevant sections for AGENTS.md |
Process
Detect source files:
ls -la | grep -E '(AGENT\.md|\.cursorrules|\.windsurfrules|CLAUDE\.md)'If no source specified, auto-detect:
- Priority: AGENT.md > .cursorrules > .windsurfrules > CLAUDE.md
- Prompt user if multiple files found
Read and analyze source file:
- Identify existing sections
- Extract commands, conventions, and guidelines
- Note any missing recommended sections
Transform to compact AGENTS.md format:
- Map existing content to recommended sections
- Add missing sections based on project analysis
- Keep core AGENTS.md lean and actionable (target: 50-80 lines)
- Focus on essential commands and conventions
- Use relative paths for references (e.g.,
./docs/agents/testing.md) - For monorepos: Add relative path references to sibling packages (e.g.,
../api/AGENTS.md)
Create the new AGENTS.md:
- Write to ./AGENTS.md (or specified path)
- Do not overwrite existing AGENTS.md without confirmation
Ask about references folder using AskUserQuestion (only if source was verbose or user wants detailed docs):
questions: [ { "question": "Would you like me to create a references folder for the detailed content from the original file?", "header": "References", "options": [ { "label": "No, single file is enough", "description": "Keep AGENTS.md as a simple, compact file. Recommended for most projects." }, { "label": "Yes, create docs/agents/", "description": "Create references folder and move detailed content there." }, { "label": "Yes, use custom path", "description": "Specify a custom path for the references folder." } ] } ]Only ask if:
- Original source file was verbose (>100 lines)
- User explicitly wants to preserve all detailed content
- Project has complex patterns needing documentation
Create reference files (if user chose yes):
- Create detailed docs in references folder (e.g., docs/agents/)
- Reference them from AGENTS.md
Validate the migrated file:
- Check all commands are executable
- Ensure all recommended sections present
- Confirm no content lost in migration
- Verify file is compact (under 100 lines)
Handle original file:
Call AskUserQuestion:
{ "questions": [{ "question": "What should I do with the original source file?", "header": "Original file", "options": [ {"label": "Create symlink", "description": "Run ln -sf AGENTS.md <original-file> so it points to the new AGENTS.md."}, {"label": "Keep as-is", "description": "Leave the original file unchanged alongside AGENTS.md."}, {"label": "Delete original", "description": "Remove the original file after migration is complete."} ] }] }- "Create symlink" →
ln -sf AGENTS.md <original-file> - "Keep as-is" → no action
- "Delete original" →
rm <original-file>
- "Create symlink" →
Ask about additional symlinks:
Call AskUserQuestion:
{ "questions": [{ "question": "Would you like me to create symlinks for other AI agents?", "header": "Symlinks", "options": [ {"label": "No symlinks", "description": "Keep AGENTS.md as the only file."}, {"label": "Create CLAUDE.md", "description": "Symlink CLAUDE.md → AGENTS.md for Claude Code compatibility."}, {"label": "Create all symlinks", "description": "Create CLAUDE.md, .cursorrules, and .windsurfrules all pointing to AGENTS.md."} ] }] }- "No symlinks" → Done
- "Create CLAUDE.md" →
ln -sf AGENTS.md CLAUDE.md - "Create all symlinks" →
ln -sf AGENTS.md CLAUDE.md && ln -sf AGENTS.md .cursorrules && ln -sf AGENTS.md .windsurfrules
Content Mapping
Transform common patterns:
| Source Pattern | AGENTS.md Section |
|---|---|
| "Run tests with..." | Testing instructions |
| "Code style: ..." | Code style |
| "Before committing..." | PR instructions |
| "Environment setup..." | Setup commands |
| "Security notes..." | Security |
Example Transformation
# Before (verbose .cursorrules, 80 lines)
## Code Style
- Use TypeScript strict mode
- Single quotes for strings
- No semicolons
- Max line length: 100 characters
- Functional components with hooks
- Colocate tests next to components
- Use meaningful variable names
- Avoid any type
- Prefer const over let
- Document complex functions with JSDoc
[... 50 more lines of detailed conventions ...]
# After (compact AGENTS.md, 60 lines)
## Code style
- TypeScript strict mode
- Single quotes, no semicolons
- Functional components with hooks
## PR instructions
- Use Conventional Commits (see agd:conventional-commits skill)
- Run `pnpm lint && pnpm test` before pushing
Validation Checklist
After migration, verify:
- All commands are executable
- All recommended sections present (or intentionally omitted)
- Original content preserved (nothing lost)
- Format follows skill best practices
- File is concise (ideally under 100 lines)
- References folder only created if user wanted it
Notes
- Always load the agents-md skill for context
- Default to compact: Create a simple AGENTS.md first
- Preserve all original content - migration should not lose information
- Only ask about references folder for verbose sources
- Ask user before deleting original files
- Validate migrated file before completing
- Offer symlink creation for backward compatibility
Source: Christophe1997/agent-extentions — distributed by TomeVault.
先判断是否适合
作者设计意图
作者的方法与取舍
边界和复核