Agent skill

monorepo-management

Monorepo tooling, task orchestration, and workspace architecture for JavaScript/TypeScript repositories. Use when setting up Turborepo, Nx, pnpm workspaces, or npm workspaces; designing package boundaries; configuring remote caching; optimizing CI for affected packages; managing versioning with Changesets; or untangling circular dependencies. Activate on "monorepo", "turborepo", "nx", "pnpm workspace", "task pipeline", "remote cache", "changesets", "CODEOWNERS", "circular dependency", "affected packages", "workspace". NOT for git submodules or multi-repo federation strategies, non-JavaScript monorepos (Bazel, Pants, Buck), or single-package repository setup.

Stars 81
Forks 12

Install this agent skill to your Project

npx add-skill https://github.com/curiositech/some_claude_skills/tree/main/.claude/skills/monorepo-management

Metadata

Additional technical details for this skill

tags
monorepo management turborepo nx
category
DevOps & Site Reliability
pairs with
[
    {
        "skill": "dependency-management",
        "reason": "Workspace-level dependency hoisting and version alignment are monorepo-specific challenges"
    },
    {
        "skill": "git-workflow-expert",
        "reason": "Monorepo git strategies (sparse checkout, path-based triggers) differ from single-repo workflows"
    },
    {
        "skill": "github-actions-pipeline-builder",
        "reason": "Monorepo CI requires path-filtered workflows and build caching across workspaces"
    },
    {
        "skill": "code-architecture",
        "reason": "Module boundaries and dependency direction rules define monorepo workspace structure"
    }
]

SKILL.md

Monorepo Management

Monorepo tooling and workspace architecture expert for JavaScript and TypeScript projects. Covers tool selection, task pipeline configuration, dependency graph management, CI optimization, and versioning strategy.

When to Use

Use for:

  • Choosing between Turborepo, Nx, Lerna, and Rush for a new or migrating repository
  • Configuring turbo.json task pipelines, remote caching, and Docker pruning
  • Structuring workspace packages: apps, packages, shared configs, internal libraries
  • Setting up pnpm or npm workspaces with correct hoisting rules
  • Eliminating circular dependencies between workspace packages
  • Configuring path-based CI with affected package detection
  • Managing package versioning and changelogs with Changesets

NOT for:

  • Git submodules or polyrepo coordination → discuss trade-offs separately
  • Bazel, Pants, or Buck for non-JavaScript monorepos
  • Single-package repository setup → use project-level tooling skills
  • Container orchestration of services within the monorepo → use docker-containerization

Tool Selection Decision Tree

mermaid
flowchart TD
    A[Monorepo tool needed] --> B{Team size and complexity?}
    B -->|Small team, apps-first| C{Need plugin ecosystem?}
    B -->|Large org, many teams| D[Rush or Nx]
    C -->|No — just fast builds| E[Turborepo]
    C -->|Yes — code gen, generators| F[Nx]
    D --> G{Microsoft/enterprise patterns?}
    G -->|Yes| H[Rush]
    G -->|No| I[Nx]
    E --> J{Package manager preference?}
    J -->|pnpm — recommended| K[pnpm + Turborepo]
    J -->|npm or yarn| L[npm/yarn workspaces + Turborepo]
    I --> M[Nx Cloud for remote caching]
    H --> N[Rush's own cache]
    K --> O[Vercel Remote Cache\nor self-hosted]

Tool Comparison Summary

Tool Best For Remote Cache Learning Curve
Turborepo Apps-first, fast builds, simple config Vercel / self-hosted Low
Nx Library-heavy, code generation, plugin ecosystem Nx Cloud / self-hosted Medium
Rush Enterprise, Microsoft stack, strict isolation Custom High
Lerna Legacy — migrating from None (use with Turborepo) Low

Recommendation for new projects: Start with pnpm workspaces + Turborepo. Migrate to Nx if you need advanced code generation or project graph visualization.


Workspace Architecture Decision Tree

mermaid
flowchart TD
    A[New package in monorepo?] --> B{Who consumes it?}
    B -->|Internal only, not published| C[Internal package:\nno versioning, workspace: protocol]
    B -->|Published to npm| D[Published package:\nChangesets, semver, CHANGELOG]
    B -->|Shared config only| E[Config package:\neslint-config-*, tsconfig-*]
    C --> F{What type?}
    F -->|UI components| G[packages/ui]
    F -->|Business logic / utilities| H[packages/utils or packages/core]
    F -->|Shared types| I[packages/types]
    F -->|API client| J[packages/api-client]
    D --> K[packages/publishable-name]
    A --> L{Is it an application?}
    L -->|Yes| M[apps/ directory:\nnext-app, api, docs-site]

Dependency Graph and Build Pipeline

mermaid
graph LR
    A[apps/web] --> B[packages/ui]
    A --> C[packages/utils]
    A --> D[packages/types]
    E[apps/api] --> C
    E --> D
    B --> D
    F[packages/ui-icons] --> D
    B --> F

    style A fill:#4a9d9e,color:#fff
    style E fill:#4a9d9e,color:#fff
    style B fill:#6b7280,color:#fff
    style C fill:#6b7280,color:#fff
    style D fill:#9ca3af
    style F fill:#9ca3af

Reading the graph: Build order flows from dependencies to dependents. packages/types (no deps) builds first. packages/ui builds after types. apps/web builds last. Turborepo and Nx compute this graph automatically — you define task dependencies, they order execution.


Turborepo Configuration

turbo.json — Task Pipeline

json
{
  "$schema": "https://turbo.build/schema.json",
  "tasks": {
    "build": {
      "dependsOn": ["^build"],
      "inputs": ["src/**/*.tsx", "src/**/*.ts", "package.json", "tsconfig.json"],
      "outputs": [".next/**", "!.next/cache/**", "dist/**"]
    },
    "test": {
      "dependsOn": ["^build"],
      "inputs": ["src/**/*.ts", "src/**/*.tsx", "**/*.test.ts", "**/*.test.tsx"],
      "outputs": ["coverage/**"]
    },
    "lint": {
      "inputs": ["src/**/*.ts", "src/**/*.tsx", ".eslintrc*"]
    },
    "typecheck": {
      "dependsOn": ["^build"]
    },
    "dev": {
      "cache": false,
      "persistent": true
    }
  }
}

Key concepts:

  • ^build means "build all dependencies first before building this package"
  • inputs determines cache keys — changes outside inputs don't invalidate the cache
  • outputs are stored in cache and restored on cache hit
  • cache: false for long-running tasks (dev servers, watchers)
  • persistent: true for tasks that don't exit

Running Tasks

bash
# Run build for all packages
turbo build

# Run only for packages affected by changes since main branch
turbo build --filter=...[origin/main]

# Run for a specific app and its dependencies
turbo build --filter=web...

# Run in parallel across packages
turbo lint typecheck --parallel

# Dry run to see what would execute
turbo build --dry-run

Remote Caching

bash
# Login to Vercel remote cache (free for open source)
npx turbo login

# Link to team/project
npx turbo link

# CI: pass token via environment
TURBO_TOKEN=$TURBO_TOKEN turbo build

Self-hosted alternative: turbo-remote-cache package or Turborepo's built-in HTTP cache server in Turborepo 2.x.

Consult references/turborepo-patterns.md for Docker pruning for deployment, scoped filtering, and advanced pipeline patterns.


pnpm Workspaces

pnpm-workspace.yaml

yaml
packages:
  - 'apps/*'
  - 'packages/*'
  - 'tools/*'

package.json workspace dependencies

json
{
  "dependencies": {
    "@myorg/ui": "workspace:*",
    "@myorg/utils": "workspace:^1.0.0"
  }
}

Use workspace:* for internal packages that should always match the local version. Use workspace:^ only for internal packages that are also published and where you want semver range resolution.

Hoisting Control

# .npmrc — control hoisting behavior
hoist=true
public-hoist-pattern[]=*eslint*
public-hoist-pattern[]=*prettier*
shamefully-hoist=false  # never — breaks encapsulation

shamefully-hoist=true is a trap: it makes all packages available everywhere but breaks encapsulation. If your tools require it, fix the tool dependency instead.


Changesets for Versioning

bash
# Initialize in your monorepo
npx changeset init

# Create a changeset when making a change
npx changeset
# Prompts: which packages changed, major/minor/patch, description

# Preview what versions will be bumped
npx changeset status

# Bump versions and update changelogs (CI or release branch)
npx changeset version

# Publish to npm
npx changeset publish

.changeset/config.json

json
{
  "$schema": "https://unpkg.com/@changesets/config@3.0.0/schema.json",
  "changelog": "@changesets/cli/changelog",
  "commit": false,
  "fixed": [],
  "linked": [],
  "access": "restricted",
  "baseBranch": "main",
  "updateInternalDependencies": "patch",
  "ignore": ["@myorg/app-web", "@myorg/app-api"]
}

Set access: "public" for open-source packages. ignore lists apps that should not be published to npm.

Consult references/workspace-architecture.md for full CODEOWNERS setup, ESLint config sharing, and shared tsconfig patterns.


Path-Based CI (Only Test What Changed)

yaml
# .github/workflows/ci.yml
name: CI

on:
  push:
    branches: [main]
  pull_request:

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0  # required for turbo --filter to work correctly

      - uses: pnpm/action-setup@v4
        with:
          version: 9

      - uses: actions/setup-node@v4
        with:
          node-version: 20
          cache: 'pnpm'

      - run: pnpm install --frozen-lockfile

      - name: Build affected packages
        run: pnpm turbo build --filter=...[origin/main]
        env:
          TURBO_TOKEN: ${{ secrets.TURBO_TOKEN }}
          TURBO_TEAM: ${{ vars.TURBO_TEAM }}

      - name: Test affected packages
        run: pnpm turbo test --filter=...[origin/main]

--filter=...[origin/main] means "run for packages whose files changed compared to main, plus all packages that depend on them (upstream consumers)."


Anti-Patterns

Anti-Pattern: No Task Caching (Rebuilding Everything)

Novice: "We run turbo build and it always rebuilds all 40 packages, even when only one changed."

Expert: Turborepo caches by computing a hash of inputs (source files + env + turbo config). If inputs haven't changed, the output is restored from cache in milliseconds. The most common cause of cache misses is missing inputs declarations — Turborepo falls back to hashing the entire package directory, including files like .DS_Store, node_modules, and IDE configs that change constantly.

Detection: Run turbo build --verbosity=2 and look for "MISS" next to packages. Check whether .turbo cache entries exist and if the hash matches between runs.

Fix: Explicitly declare inputs in turbo.json to include only files that affect build output. Explicitly declare outputs so the cache knows what to store. Add non-source files to .gitignore and .turboignore.

LLM mistake: Tutorials often omit inputs and outputs because a working demo doesn't need them. Production repos require explicit declarations or cache hit rates stay near 0%.


Anti-Pattern: Circular Dependencies Between Workspace Packages

Novice: "packages/auth imports from packages/api-client and packages/api-client imports from packages/auth — is that a problem?"

Expert: Yes. Circular dependencies make build order impossible to determine. Turborepo will error on cycles. More importantly, circular deps indicate a design flaw: two packages whose concerns are entangled. The fix is to extract the shared types or utilities to a third package that both can import without creating a cycle.

Detection:

bash
# Turborepo detects cycles and refuses to run
turbo build  # "Error: Package graph cycle detected"

# Manual detection with madge
npx madge --circular --extensions ts packages/

Fix:

Before:
  packages/auth → packages/api-client → packages/auth (cycle!)

After:
  packages/types (new: shared auth types, no dependencies)
  packages/api-client → packages/types
  packages/auth → packages/types
  packages/auth → packages/api-client (one-way, no cycle)

Timeline: This is not a new problem — circular deps have been a JavaScript packaging issue since npm v1 (2010). The reason it appears in monorepos specifically is that workspace packages make it easy to import across package boundaries without thinking about dependency direction.


Anti-Pattern: Using shamefully-hoist=true in pnpm

Novice: "My CLI tool can't find its peer dependency. I'll add shamefully-hoist=true to .npmrc to fix it."

Expert: shamefully-hoist makes pnpm behave like npm/yarn classic, putting all packages in a flat node_modules. This "fixes" the immediate issue but breaks pnpm's strict isolation, which is the entire reason to use pnpm. The right fix is to add the missing peer dependency to the package that needs it, or use public-hoist-pattern to hoist only the specific package that requires it.

Detection: Any .npmrc with shamefully-hoist=true in a pnpm workspace.


References

  • references/turborepo-patterns.md — Consult when configuring turbo.json, setting up remote caching, using Docker pruning for deployment, or debugging cache misses.
  • references/workspace-architecture.md — Consult when designing package boundaries, sharing ESLint/TypeScript configs, setting up CODEOWNERS, or planning a migration from single-repo to monorepo.

Expand your agent's capabilities with these related and highly-rated skills.

curiositech/some_claude_skills

3d-cv-labeling-2026

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

81 12
Explore
curiositech/some_claude_skills

project-management-guru-adhd

Expert project manager for ADHD engineers managing multiple concurrent projects. Specializes in hyperfocus management, context-switching minimization, and parakeet-style gentle reminders. Activate on 'ADHD project management', 'context switching', 'hyperfocus', 'task prioritization', 'multiple projects', 'productivity for ADHD', 'task chunking', 'deadline management'. NOT for neurotypical project management, rigid waterfall processes, or general productivity advice without ADHD context.

81 12
Explore
curiositech/some_claude_skills

large-scale-map-visualization

Master of high-performance web map implementations handling 5,000-100,000+ geographic data points. Specializes in Leaflet.js optimization, Supercluster algorithms, viewport-based loading, canvas rendering, and progressive disclosure UX patterns.

81 12
Explore
curiositech/some_claude_skills

adhd-design-expert

Designs digital experiences for ADHD brains using neuroscience research and UX principles. Expert in reducing cognitive load, time blindness solutions, dopamine-driven engagement, and compassionate design patterns. Activate on 'ADHD design', 'cognitive load', 'accessibility', 'neurodivergent UX', 'time blindness', 'dopamine-driven', 'executive function'. NOT for general accessibility (WCAG only), neurotypical UX design, or simple UI styling without ADHD context.

81 12
Explore
curiositech/some_claude_skills

liaison

Translate multi-agent ecosystem activity into human-readable status briefings, decision requests, and progress summaries. Use for 'status update', 'brief me', 'what happened', 'summarize progress'. NOT for project planning (use project-management-guru-adhd), code review, or technical documentation.

81 12
Explore
curiositech/some_claude_skills

windows-95-web-designer

Modern web applications with authentic Windows 95 aesthetic. Gradient title bars, Start menu paradigm, taskbar patterns, 3D beveled chrome. Extrapolates Win95 to AI chatbots, mobile UIs, responsive layouts. Activate on 'windows 95', 'win95', 'start menu', 'taskbar', 'retro desktop', '95 aesthetic', 'clippy'. NOT for Windows 3.1 (use windows-3-1-web-designer), vaporwave/synthwave, macOS, flat design.

81 12
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results