By xiqin
Orchestrate AI-assisted engineering workflows with a structured pipeline framework: TDD test-first cycles, parallel subagent task execution, QA acceptance pipelines, code review triage and request generation, isolated git worktree development, and housekeeping that syncs memory and indexes. Supports ambiguous-request routing and multi-skill dispatch.
Explore 2-3 implementation options with trade-offs when the user describes a new feature or requirement. Use when: the user asks for a new feature design, implementation options, or technical trade-off analysis.
Dispatch multiple independent tasks to parallel subagents when no shared file conflicts exist. Use when: a confirmed plan has independent tasks that can run concurrently without touching shared files.
Clean up a feature branch after verification: merge, create PR, keep, or discard. Follows conventional commits. Use when: a development branch has passed verification and needs merge, PR, keep, or discard handling.
Synchronize graph backend index, structured memory, and entry docs with code after verification passes. Use when: verified code changes need graph index sync, memory updates, or entry documentation refresh.
Bootstrap .loom/ context files for a new repository: constitution, structured memory, workflow, and agent entry files. Use when: initializing loom context for a repository that lacks .loom/ files or standard agent entry docs.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Executables (bin/) — files in this plugin's bin/directory are added to the Bash tool's PATH while the plugin is enabled.
AI 工程化框架。把需求、规范、上下文、执行过程"织"成一套稳定工程流程。
feature / bugfix / hotfix / refactor / quickfix / chore / qa / pm-prototype 自动选择对应步骤loom),负责安装、更新、诊断、卸载,以及流水线执行(run/status/tasks/index)、证据导出(evidence)、结构化记忆(memory)和 MCP server(mcp-serve)/loom-init-project 或 loom init-project),自动扫描项目生成宪章、结构化记忆和 agent 入口从需求描述出发,经过头脑风暴、计划拆解、隔离开发、代码审查,最终交付。
| 工具 | 支持等级 | 入口文件 | Skills | Hooks | Plugin | MCP 配置 |
|---|---|---|---|---|---|---|
| Claude Code | full | CLAUDE.md | ✅ | ✅ | ✅ | ✅ |
| Codex | full | AGENTS.md | ✅ | ✗ | ✗ | ✅ |
| Cursor | full | .cursor/rules/*.mdc | ✅ | ✗ | ✗ | ✅ |
| GitHub Copilot | full | .github/copilot-instructions.md | ✅ | ✗ | ✗ | ✗ |
| OpenCode | full | AGENTS.md | ✅ | ✅ | ✅ | ✅ |
config/tools.schema.json 中仍标记 hooksSupport: false,后续生命周期扩展见 docs/evolution-roadmap.mdconfig/tools.schema.json 的 contract 为单一源,并生成到 src/generated/tooling.js 的 ADAPTER_CONTRACTS;loom doctor 会基于契约输出一致性和版本诊断loom index 自动委派给它做 AST 级图索引,并注册其 MCP server;未装则跳过图查询能力git clone https://github.com/xiqin/loom.git
cd loom
bash install.sh --tool claude-code
远程一键安装:
# Unix
curl -fsSL https://raw.githubusercontent.com/xiqin/loom/main/install.sh | bash -s -- --tool claude-code
# Windows PowerShell
irm https://raw.githubusercontent.com/xiqin/loom/main/install.ps1 -OutFile install.ps1; .\install.ps1 -Tool claude-code
npm i -g loom-engineering
loom install --tool claude-code
| Flag | 作用 |
|---|---|
--tool <targets> | 目标工具(必填,逗号分隔或 "all") |
--dry-run | 预览,不实际写入 |
--from-release | 从 GitHub release tag 下载(可重现安装) |
--version <ver> | 指定下载版本(配合 --from-release) |
loom doctor # 诊断安装状态
loom list # 列出可用 skills 和 commands
loom 的 MCP server 支持通过环境变量控制工具暴露和运行统计。需要把变量配置到客户端启动 loom mcp-serve 的 MCP server 配置里,而不是只在普通 shell 会话里临时设置。
| 变量 | 建议值 | 作用 |
|---|---|---|
LOOM_LAZY_TOOLS | 1 | 启用 MCP 工具懒加载,默认只暴露基础工具,按需加载 pipeline/context/memory/session 工具组,减少工具定义带来的上下文开销。 |
LOOM_TELEMETRY | 1 | 开启 loom MCP 工具调用统计,记录调用次数、耗时、响应字节数和估算响应 token;可通过 loom_telemetry 工具查看,MCP server 退出时也会向 stderr 输出摘要。 |
配置示例:
Claude Code (~/.claude.json):
{
"mcpServers": {
"loom": {
"command": "loom",
"args": ["mcp-serve"],
"env": {
"LOOM_LAZY_TOOLS": "1",
"LOOM_TELEMETRY": "1",
},
},
},
}
// OpenCode: mcp.loom.environment
{
"mcp": {
"loom": {
"type": "local",
"command": ["loom", "mcp-serve"],
"enabled": true,
"environment": {
"LOOM_LAZY_TOOLS": "1",
"LOOM_TELEMETRY": "1",
},
},
},
}
修改 MCP 配置后需要重启对应客户端,运行中的 MCP server 不会自动继承新环境变量。
在 Codex/Claude/OpenCode 中直接触发 loom-init-project skill 即可;脚本由 skill 自动运行,不需要用户手动调用。
也可以使用 CLI:
loom init-project
# 脚本卸载
bash uninstall.sh --tool claude-code
.\uninstall.ps1 -Tool claude-code
# CLI 卸载
loom uninstall --tool claude-code
卸载只清理用户级安装的文件(用户目录下的 skills、commands、plugin 注册),不碰项目目录中的任何文件。
如果误卸载,重新安装即可:
loom install --tool claude-code
package.json 中定义,通过 scripts/sync-version.mjs 同步到所有元数据文件loom:version=x.y.z 标记,用于检测已安装版本loom update 自动比较版本号,仅在版本不同时更新loom doctor 显示当前安装状态和版本loom doctor
# 输出示例:
# Tool: claude-code
# Version: 2.0.1
# Status: installed
流水线由 .loom/workflow.yaml 集中定义。支持两种模式:
npx claudepluginhub xiqin/loom --plugin loomStandalone image generation plugin using Nano Banana MCP server. Generates and edits images, icons, diagrams, patterns, and visual assets via Gemini image models. No Gemini CLI dependency required.
Frontend design taste skills including brutalist, minimalist, soft, redesign, stitch, and more
Unified capability management center for Skills, Agents, and Commands.
Ultra-compressed communication mode. Cuts 65% of output tokens (measured) while keeping full technical accuracy by speaking like a caveman.
Frontend design skill for UI/UX implementation
Memory compression system for Claude Code - persist context across sessions