devsync-mcp
MCP server for AI-merge of team dev configurations, allowing the IDE's LLM to intelligently merge team configs with personal configs.
README
devsync-mcp
MCP server for AI-merge of team dev configurations. Works with any MCP-capable IDE (Claude Code, Cursor, Windsurf, Copilot, Zed, etc.).
What it does
Teams maintain a shared config profile repo (shell configs, git settings, AI rules, editor configs). devsync-mcp provides MCP tools that let the IDE's LLM intelligently merge team configs with your personal configs — preserving your customizations while incorporating team standards.
Key insight: The MCP server does NOT call an LLM. It provides source and target content to the host IDE's LLM, which performs the intelligent merge.
Install
pip install devsync-mcp
MCP Configuration
Add to your IDE's MCP settings:
{
"mcpServers": {
"devsync-mcp": {
"command": "devsync-mcp",
"transport": "stdio"
}
}
}
Tools
| Tool | Purpose |
|---|---|
pull_team_profile |
Clone/pull team config repo |
list_profile_configs |
List configs in a profile |
detect_current_configs |
Scan environment for existing configs |
preview_merge |
Return source + target for LLM merge |
apply_merge |
Write merged content with backup |
sync_all |
Pull latest + present merge plan |
list_backups |
List previous merge backups |
restore_backup |
Restore from backup |
get_merge_status |
Show merged/pending/changed status |
AI-Merge Flow
1. pull_team_profile(git_url) → downloads team config repo
2. list_profile_configs(profile) → shows available configs
3. detect_current_configs() → finds your existing configs
4. preview_merge(profile, config) → returns {source_content, target_content}
5. HOST LLM reads both, produces intelligent merge
6. apply_merge(target, merged) → writes with backup
Team Profile Format
Create a devsync-profile.yaml in your team config repo:
name: acme-team-config
description: ACME Corp standard dev environment
version: 1.0.0
configs:
- name: zshrc-additions
file: shell/zshrc-additions.sh
target: ~/.zshrc
type: shell_profile
description: Team shell aliases and PATH additions
- name: claude-rules
file: ai-rules/claude-rules.md
target: CLAUDE.md
type: ai_rules
description: Team coding standards for Claude Code
Config Types
shell_profile—.zshrc,.bashrc,.bash_profile,.profilegit_config—.gitconfig,.gitignore_globaleditor_config—.editorconfigai_rules—CLAUDE.md,.cursorrules,AGENTS.md, etc.vscode_settings—.vscode/settings.json, etc.ssh_config—.ssh/configcustom— any file with explicit source/target mapping
Claude Code Skill
Install the /sync-team skill for a guided experience:
cp -r skill/sync-team ~/.claude/skills/
Then use /sync-team in Claude Code to walk through the full sync flow.
Development
git clone https://github.com/troylar/devsync-mcp
cd devsync-mcp
pip install -e ".[dev]"
pytest
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。