windsurf-architecture-variants
Choose effective workspace structures for Windsurf based on team size and project complexity.
Install
mkdir -p .claude/skills/windsurf-architecture-variants && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/9123" && unzip -o skill.zip -d .claude/skills/windsurf-architecture-variants && rm skill.zipInstalls to .claude/skills/windsurf-architecture-variants
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Choose workspace architectures for different project scales in Windsurf.Key capabilities
- →Structure monorepos for optimal workspace indexing
- →Configure per-service .windsurfrules and .codeiumignore
- →Partition large codebases into focused workspace windows
- →Implement polyglot project layouts
- →Optimize Cascade context for team collaboration
How it works
It provides blueprints for workspace partitioning based on project size, ensuring Cascade indexes only relevant files to maintain performance and context quality.
Inputs & outputs
When to use windsurf-architecture-variants
- →Structure a new monorepo for Windsurf
- →Optimize workspace indexing for performance
- →Configure project layout for team collaboration
- →Plan partitioning for large codebases
About this skill
Windsurf Architecture Variants
Overview
How you structure your Windsurf workspace directly impacts Cascade's effectiveness. Large monorepos, multi-service setups, polyglot codebases, and different team sizes each require different approaches. This skill covers workspace strategies from solo projects to 100+ developer organizations.
Prerequisites
- Windsurf installed
- Understanding of Cascade's workspace indexing model
- Git workflow established
Instructions
Variant 1: Single Project (Solo / Small Team)
Best for: 1-3 developers, single service, <10K files.
my-project/
├── .windsurfrules # Full project context
├── .codeiumignore # Exclude build artifacts
├── src/
├── tests/
├── package.json
└── README.md
Configuration:
- Open entire project as workspace
- Cascade indexes everything — no partitioning needed
.windsurfrulescontains complete stack and architecture details
Variant 2: Focused Monorepo Windows (Medium Team)
Best for: 3-15 developers, monorepo with 2-10 packages.
monorepo/
├── .windsurfrules # Brief shared conventions
├── .codeiumignore # Aggressive exclusions at root
├── packages/
│ ├── api/
│ │ ├── .windsurfrules # API-specific rules
│ │ └── .codeiumignore
│ ├── web/
│ │ ├── .windsurfrules # Frontend-specific rules
│ │ └── .codeiumignore
│ └── shared/
│ ├── .windsurfrules # Library conventions
│ └── .codeiumignore
└── .windsurf/
└── workflows/ # Shared workflows
Strategy:
# Each developer opens their package directory:
windsurf packages/api/ # Backend dev
windsurf packages/web/ # Frontend dev
windsurf packages/shared/ # Library maintainer
# NOT: windsurf monorepo/ # Too broad!
Variant 3: Multi-Window Team Workflow (Large Team)
Best for: 15+ developers, microservices, 50K+ total files.
Developer A: Windsurf → services/auth/ (auth service)
Developer B: Windsurf → services/payments/ (payments)
Developer C: Windsurf → services/notifications/ (notifications)
Developer D: Windsurf → shared/libs/ (shared libraries)
Each developer gets focused Cascade context per workspace window.
Team conventions:
1. One Windsurf window per service/package
2. Every service has its own .windsurfrules and .codeiumignore
3. Cascade tasks scoped to current workspace only
4. Cross-service changes: open both workspaces side by side
5. Tag cascade commits: git commit -m "[cascade] description"
6. Use shared workflows from central config repo
Variant 4: Polyglot / Multi-Language
Best for: Projects with multiple languages (TypeScript + Python + Go).
# Each language has different .windsurfrules
services/
├── ts-api/
│ └── .windsurfrules # TypeScript patterns, Fastify, Vitest
├── python-ml/
│ └── .windsurfrules # Python patterns, FastAPI, pytest
└── go-gateway/
└── .windsurfrules # Go patterns, chi router, go test
<!-- .windsurfrules for Python service -->
# Project: ML Pipeline
## Stack
- Language: Python 3.11
- Framework: FastAPI
- ML: scikit-learn, pandas
- Testing: pytest with fixtures
- Type checking: mypy (strict)
## Conventions
- Use pydantic for all data models
- Async endpoints with asyncio
- Type hints on all functions
- No print() — use logging module
Variant 5: Frontend-Heavy (Design System)
Best for: UI-heavy projects with design system, Storybook, component library.
<!-- .windsurfrules for design system -->
# Project: Design System
## Stack
- Framework: React 18 + Next.js 14
- Styling: Tailwind CSS + custom tokens
- Components: Radix UI primitives
- Docs: Storybook 8
- Testing: Vitest + Testing Library
## Component Conventions
- One component per file (ComponentName.tsx)
- Co-located tests: ComponentName.test.tsx
- Co-located stories: ComponentName.stories.tsx
- Props interface exported: ComponentNameProps
- Use forwardRef for all components
- Use CVA (class-variance-authority) for variants
## Design Tokens
- Colors: use design-system/tokens, never raw Tailwind colors
- Spacing: use space-* scale (4px base)
- Typography: use text-* presets
Cascade integration: Use Previews to iterate on UI components:
"Preview the Button component with all variants"
Click elements in Preview → send to Cascade for refinement
Decision Matrix
| Factor | Solo | Focused Mono | Multi-Window | Polyglot |
|---|---|---|---|---|
| Team Size | 1-3 | 3-15 | 15+ | Any |
| Codebase | <10K files | 10K-50K | 50K+ | Mixed |
| Cascade Speed | Fast | Fast (per window) | Fast (per window) | Fast (per window) |
| Setup Effort | Minimal | .codeiumignore + rules | Per-service config | Per-language rules |
| Context Quality | Excellent | Good | Good | Good (per lang) |
Error Handling
| Issue | Cause | Solution |
|---|---|---|
| Cascade is slow | Too many files indexed | Open smaller workspace, add .codeiumignore |
| Wrong file context | Monorepo root open | Open specific service directory |
| Conflicting edits | Multiple devs, same files | Feature branches per Cascade session |
| Wrong language patterns | Multi-language workspace | Separate .windsurfrules per language directory |
| Stale suggestions | Index out of date | Command Palette > "Codeium: Reset Indexing" |
Examples
Optimized .codeiumignore (Universal)
node_modules/
dist/
build/
.next/
coverage/
*.min.js
*.map
__pycache__/
.venv/
target/
vendor/
*.log
*.sqlite
Workspace Health Check
set -euo pipefail
FILE_COUNT=$(find . -type f -not -path '*/node_modules/*' -not -path '*/.git/*' | wc -l)
echo "Indexed files: ~$FILE_COUNT"
[ "$FILE_COUNT" -gt 10000 ] && echo "WARNING: Consider opening a subdirectory"
[ -f .windsurfrules ] && echo "Rules: $(wc -c < .windsurfrules) chars" || echo "Rules: MISSING"
[ -f .codeiumignore ] && echo "Ignore: $(wc -l < .codeiumignore) patterns" || echo "Ignore: MISSING"
Resources
Next Steps
For known pitfalls and anti-patterns, see windsurf-known-pitfalls.
Prerequisites
Limitations
- →Requires manual configuration of per-service rules
- →Large monorepos require careful partitioning to avoid indexing delays
How it compares
It offers specific architectural variants for different team sizes and project types rather than a one-size-fits-all workspace approach.
Compared to similar skills
windsurf-architecture-variants side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| windsurf-architecture-variants (this skill) | 0 | 25d | Review | Intermediate |
| software-architecture | 333 | 6mo | No flags | Intermediate |
| architect-review | 109 | 4mo | No flags | Advanced |
| mcp-builder | 136 | 3mo | Review | Advanced |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by jeremylongshore
View all by jeremylongshore →You might also like
software-architecture
davila7
Guide for quality focused software architecture. This skill should be used when users want to write code, design architecture, analyze code, in any case that relates to software development.
architect-review
sickn33
Master software architect specializing in modern architecture patterns, clean architecture, microservices, event-driven systems, and DDD. Reviews system designs and code changes for architectural integrity, scalability, and maintainability. Use PROACTIVELY for architectural decisions.
mcp-builder
anthropics
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
solid-principles
SmidigStorm
Enforce SOLID principles (Single Responsibility, Open/Closed, Liskov Substitution, Interface Segregation, Dependency Inversion) in object-oriented design. Use when writing or reviewing classes and modules.
codex
Lucklyric
Invoke Codex CLI for complex coding tasks requiring high reasoning capabilities. This skill should be invoked when users explicitly mention "Codex", request complex implementation challenges, advanced reasoning, or need high-reasoning model assistance. Automatically triggers on codex-related requests and supports session continuation for iterative development.
architecture-patterns
wshobson
Implement proven backend architecture patterns including Clean Architecture, Hexagonal Architecture, and Domain-Driven Design. Use when architecting complex backend systems or refactoring existing applications for better maintainability.