Orchestrates the full lifecycle of a feature implementation from plan to verification.

Install

mkdir -p .claude/skills/feature-tanjed && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/14367" && unzip -o skill.zip -d .claude/skills/feature-tanjed && rm skill.zip

Installs to .claude/skills/feature-tanjed

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.

Implement a complete feature end-to-end: plan, scaffold structure, write all layers of code, and write tests. Auto-invoke when the user says "build the X feature", "implement the full flow for Y", "end-to-end implementation of Z", or "I need everything for the order cancellation feature".
289 chars✓ has a “when” triggerlonger than Claude Code's old 250-char listing cap (fine on current versions)
Advanced

Key capabilities

  • Plan a complete feature implementation
  • Scaffold directory structures for new features
  • Write code for all layers of a feature
  • Generate unit and integration tests
  • Self-verify code with build and test commands

How it works

The skill orchestrates a multi-phase development cycle, starting with planning and context reading, then implementing code layers, and finally generating and running tests.

Inputs & outputs

You give it
Feature description
You get back
Feature plan, code files, test files, and a summary report

When to use feature

  • Implement new API feature
  • Add user authentication flow
  • Build database integration

About this skill

Feature Skill

Implement a complete feature end-to-end: $ARGUMENTS

This skill orchestrates the full development cycle — design → code → test — for a self-contained feature.


Phase 1 — Understand and Plan

1.1 Read context

# Detect stack
ls package.json go.mod composer.json 2>/dev/null
 
# Understand existing patterns
find src app internal -name "*.ts" -o -name "*.go" -o -name "*.php" 2>/dev/null | head -20

Read 2–3 existing similar features/services to match the codebase's patterns exactly. Read the relevant process file for the detected stack.

1.2 Present a Feature Plan

Before writing anything, output this plan and wait for explicit approval:

Feature: <name>
 
Layers affected:
  - Controller:   <endpoint(s) it will expose>
  - Service:      <business logic — what rules does it enforce?>
  - Repository:   <what data does it read/write?>
  - Events:       <does this emit any events? what?>
 
New files:
  - <file path> — <purpose>
  - ...
 
Modified files:
  - <file path> — <what changes>
 
Interfaces / contracts to define:
  - <IXxxRepository> — methods: ...
  - <IXxxService> — methods: ...
 
DTOs:
  - <XxxRequestDTO> — fields: ...
  - <XxxResponseDTO> — fields: ...
 
Error classes:
  - <XxxNotFoundError>, <XxxAlreadyExistsError>, etc.
 
Tests to write:
  - Unit: <service> — <N> test cases
  - Integration: <repository> — <N> test cases
 
Open questions (if any):
  - <question> — assumption I'll use if you don't answer

Phase 2 — Implement (after approval)

Write files in this strict order — each layer depends on the one above:

2.1 Contracts First (no dependencies)

  • Interfaces for the service and repository
  • DTOs (input and output shapes)
  • Custom error/exception classes
  • Enums if new domain constants are needed

2.2 Repository Layer

  • Implement the repository interface against the ORM/DB
  • All DB access lives here — no raw queries in services
  • Match the DB naming conventions from ~/.claude/process/db_process.md

2.3 Service Layer

  • Inject repository via its interface
  • Enforce business rules here — validation, domain logic, event emission
  • No DB knowledge — only calls repository methods
  • Use early returns for guard clauses

2.4 Controller / Handler Layer

  • Inject service via its interface
  • Parse and validate request input (Form Request / Zod / binding)
  • Call exactly one service method per endpoint
  • Map service output to HTTP response via DTO/Resource
  • Zero business logic here

2.5 Route / DI Wiring

  • Register the route pointing to the new controller
  • Bind interfaces to implementations in the DI container / service provider
  • Never skip this — untested wiring is a common failure point

Phase 3 — Tests (spawn test-writer agent)

Spawn the test-writer agent to write in parallel:

Unit tests (service layer):

  • Mock the repository
  • Cover: happy path, each error condition, each business rule branch
  • Format: it should <behaviour> when <condition>

Integration tests (repository layer):

  • Use real test database
  • Cover: create, read, update, delete — verify data actually persists
  • Cover: constraints and edge cases (duplicate, not found, cascade)

Feature / HTTP tests (where applicable):

  • Test the full request → response cycle
  • Cover: valid input returns correct response, invalid input returns correct error code

Phase 4 — Self-Verify

# Node.js
npx tsc --noEmit && npx jest --testPathPattern="<feature>"
 
# Go
go build ./... && go test ./... -run "TestOrder"
 
# PHP
./vendor/bin/pint && ./vendor/bin/phpstan analyse && ./vendor/bin/pest --filter="order"

Fix any type errors, linting failures, or test failures before finishing.


Output Summary

After completing all phases, print:

Feature: <name> — COMPLETE
 
Files written:
  Interfaces:    <list>
  DTOs:          <list>
  Errors:        <list>
  Repository:    <list>
  Service:       <list>
  Controller:    <list>
  Routes/DI:     <list>
  Tests:         <list> (<N> unit, <N> integration, <N> feature)
 
Patterns applied: <list with one-line explanation each>
Build: PASS | FAIL (details)
Tests: <N> passing | <N> failing (details)
 
Next steps:
  - <anything that needs a follow-up decision or human action>

When not to use it

  • When only a small code change is needed
  • When the feature is not self-contained
  • When explicit approval for a feature plan is not desired

Limitations

  • Requires explicit approval for the feature plan before implementation
  • Relies on existing codebase patterns and process files for context
  • Assumes the presence of specific build and test tools for self-verification

How it compares

This skill automates the entire feature development process from design to testing, providing a structured and repeatable workflow compared to manual, ad-hoc development.

Compared to similar skills

feature side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
feature (this skill)04moReviewAdvanced
implement-feature17moReviewAdvanced
feature02moNo flagsIntermediate
add-policy01moReviewAdvanced

Try saying

Example prompts that trigger this skill in your AI assistant.

You might also like

implement-feature

tddworks

Guide for implementing features in ClaudeBar following architecture-first design, TDD, rich domain models, and Swift 6.2 patterns. Use this skill when: (1) Adding new functionality to the app (2) Creating domain models that follow user's mental model (3) Building SwiftUI views that consume domain models directly (4) User asks "how do I implement X" or "add feature Y" (5) Implementing any feature that spans Domain, Infrastructure, and App layers

16

feature

PABERTHIER

>

00

add-policy

microsoft

Use when adding, modifying, or reviewing VS Code configuration policies. Covers the full policy lifecycle from registration to export to platform-specific artifacts. Run on ANY change that adds a `policy:` field to a configuration property.

00

roo-windows-code-authoring

dejwk

roo_windows-specific addendum for code authoring and validation. Use with the local embedded-cpp-code-authoring instruction when editing public APIs, adding tests, documenting behavior, running Bazel tests, or checking example-sketch builds.

00

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.

333868

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.

32238

Search skills

Search the agent skills registry