incremental-implementation
Delivers changes incrementally in thin vertical slices. Use when implementing any feature or change that touches more than one file, when you're about to write a large amount of code at once, or when a task feels too big to land in one step. Each increment leaves the system in a working, testable state.
Best use case
incremental-implementation is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Delivers changes incrementally in thin vertical slices. Use when implementing any feature or change that touches more than one file, when you're about to write a large amount of code at once, or when a task feels too big to land in one step. Each increment leaves the system in a working, testable state.
Teams using incremental-implementation should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/incremental-implementation/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How incremental-implementation Compares
| Feature / Agent | incremental-implementation | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Delivers changes incrementally in thin vertical slices. Use when implementing any feature or change that touches more than one file, when you're about to write a large amount of code at once, or when a task feels too big to land in one step. Each increment leaves the system in a working, testable state.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Incremental Implementation
## Overview
Build in thin vertical slices — implement one piece, test it, verify it, then expand. Avoid implementing an entire feature in one pass. Each increment should leave the system in a working, testable state. This is the execution discipline that makes large features manageable.
## Instructions
### The Increment Cycle
For each slice:
1. **Implement** the smallest complete piece of functionality
2. **Test** — run the test suite (or write a test if none exists)
3. **Verify** — confirm the slice works (tests pass, build succeeds)
4. **Commit** — save progress with a descriptive message
5. **Move to the next slice**
### Slicing Strategies
**Vertical Slices (Preferred)** — Build one complete path through the stack:
```
Slice 1: Create a task (DB + API + basic UI) → User can create a task
Slice 2: List tasks (query + API + UI) → User can see their tasks
Slice 3: Edit a task (update + API + UI) → User can modify tasks
Slice 4: Delete a task (delete + API + UI) → Full CRUD complete
```
**Contract-First Slicing** — When backend and frontend develop in parallel:
```
Slice 0: Define API contract (types, interfaces, OpenAPI spec)
Slice 1a: Backend against the contract + API tests
Slice 1b: Frontend against mock data matching the contract
Slice 2: Integrate and test end-to-end
```
**Risk-First Slicing** — Tackle the riskiest piece first:
```
Slice 1: Prove the WebSocket connection works (highest risk)
Slice 2: Build real-time updates on the proven connection
Slice 3: Add offline support and reconnection
```
### Implementation Rules
**Simplicity First:** Before writing code, ask "What is the simplest thing that could work?"
```
✗ Generic EventBus with middleware pipeline for one notification
✓ Simple function call
✗ Abstract factory pattern for two similar components
✓ Two straightforward components with shared utilities
✗ Config-driven form builder for three forms
✓ Three form components
```
**Scope Discipline:** Touch only what the task requires. Don't "clean up" adjacent code, refactor unrelated imports, or add features not in the spec.
```
NOTICED BUT NOT TOUCHING:
- src/utils/format.ts has an unused import (unrelated to this task)
- The auth middleware could use better error messages (separate task)
→ Want me to create tasks for these?
```
**One Thing at a Time:** Each increment changes one logical thing. Don't mix a new component, a refactor, and a build config change in one commit.
**Keep It Compilable:** After each increment, the project must build and existing tests must pass.
**Feature Flags for Incomplete Features:**
```typescript
const ENABLE_TASK_SHARING = process.env.FEATURE_TASK_SHARING === 'true';
if (ENABLE_TASK_SHARING) {
// New sharing UI — merged but not exposed
}
```
**Rollback-Friendly:** Each increment should be independently revertable. Prefer additive changes; keep modifications minimal and focused.
## Examples
### Directing an Agent
```
"Let's implement Task 3 from the plan.
Start with just the database schema change and the API endpoint.
Don't touch the UI yet — we'll do that in the next increment.
After implementing, run `npm test` and `npm run build` to verify
nothing is broken."
```
### Increment Checklist
After each increment:
- [ ] The change does one thing and does it completely
- [ ] All existing tests still pass (`npm test`)
- [ ] The build succeeds (`npm run build`)
- [ ] Type checking passes (`npx tsc --noEmit`)
- [ ] The new functionality works as expected
- [ ] The change is committed with a descriptive message
## Guidelines
### Common Rationalizations
| Rationalization | Reality |
|---|---|
| "I'll test it all at the end" | Bugs compound. A bug in Slice 1 makes Slices 2-5 wrong |
| "It's faster to do it all at once" | Feels faster until something breaks in 500 changed lines |
| "These changes are too small to commit separately" | Small commits are free. Large commits hide bugs |
| "This refactor is small enough to include" | Refactors mixed with features make both harder to review |
### Red Flags
- More than 100 lines written without running tests
- Multiple unrelated changes in a single increment
- "Let me just quickly add this too" scope expansion
- Build or tests broken between increments
- Building abstractions before the third use case demands it
- Touching files outside the task scope "while I'm here"
### Verification
After completing all increments:
- [ ] Each increment was individually tested and committed
- [ ] The full test suite passes
- [ ] The build is clean
- [ ] The feature works end-to-end as specified
- [ ] No uncommitted changes remainRelated Skills
zustand
You are an expert in Zustand, the small, fast, and scalable state management library for React. You help developers manage global state without boilerplate using Zustand's hook-based stores, selectors for performance, middleware (persist, devtools, immer), computed values, and async actions — replacing Redux complexity with a simple, un-opinionated API in under 1KB.
zoho
Integrate and automate Zoho products. Use when a user asks to work with Zoho CRM, Zoho Books, Zoho Desk, Zoho Projects, Zoho Mail, or Zoho Creator, build custom integrations via Zoho APIs, automate workflows with Deluge scripting, sync data between Zoho apps and external systems, manage leads and deals, automate invoicing, build custom Zoho Creator apps, set up webhooks, or manage Zoho organization settings. Covers Zoho CRM, Books, Desk, Projects, Creator, and cross-product integrations.
zod
You are an expert in Zod, the TypeScript-first schema declaration and validation library. You help developers define schemas that validate data at runtime AND infer TypeScript types at compile time — eliminating the need to write types and validators separately. Used for API input validation, form validation, environment variables, config files, and any data boundary.
zipkin
Deploy and configure Zipkin for distributed tracing and request flow visualization. Use when a user needs to set up trace collection, instrument Java/Spring or other services with Zipkin, analyze service dependencies, or configure storage backends for trace data.
zig
Expert guidance for Zig, the systems programming language focused on performance, safety, and readability. Helps developers write high-performance code with compile-time evaluation, seamless C interop, no hidden control flow, and no garbage collector. Zig is used for game engines, operating systems, networking, and as a C/C++ replacement.
zed
Expert guidance for Zed, the high-performance code editor built in Rust with native collaboration, AI integration, and GPU-accelerated rendering. Helps developers configure Zed, create custom extensions, set up collaborative editing sessions, and integrate AI assistants for productive coding.
zeabur
Expert guidance for Zeabur, the cloud deployment platform that auto-detects frameworks, builds and deploys applications with zero configuration, and provides managed services like databases and message queues. Helps developers deploy full-stack applications with automatic scaling and one-click marketplace services.
zapier
Automate workflows between apps with Zapier. Use when a user asks to connect apps without code, automate repetitive tasks, sync data between services, or build no-code integrations between SaaS tools.
zabbix
Configure Zabbix for enterprise infrastructure monitoring with templates, triggers, discovery rules, and dashboards. Use when a user needs to set up Zabbix server, configure host monitoring, create custom templates, define trigger expressions, or automate host discovery and registration.
yup
Validate data with Yup schemas. Use when adding form validation, defining API request schemas, validating configuration, or building type-safe validation pipelines in JavaScript/TypeScript.
yt-dlp
Download video and audio from YouTube and other platforms with yt-dlp. Use when a user asks to download YouTube videos, extract audio from videos, download playlists, get subtitles, download specific formats or qualities, batch download, archive channels, extract metadata, embed thumbnails, download from social media platforms (Twitter, Instagram, TikTok), or build media ingestion pipelines. Covers format selection, audio extraction, playlists, subtitles, metadata, and automation.
youtube-transcription
Transcribe YouTube videos to text using OpenAI Whisper and yt-dlp. Use when the user wants to get a transcript from a YouTube video, generate subtitles, convert video speech to text, create SRT/VTT captions, or extract spoken content from YouTube URLs.