Skip to content

Repository files navigation

yolo

plan it. build it. test it. ship it. one prompt.

Stars Last Commit License

What It DoesBefore/AfterInstallUsageLanguages


An AI coding agent plugin that turns your assistant into a fully autonomous developer. Give it a task, confirm the plan, sit back. It breaks the work into tasks, executes each one, runs your linters and tests, creates git commits, and delivers a formatted report when done.

No babysitting. No "can you also run the tests?" No "don't forget to commit."

What It Does

Phase What happens
Plan Breaks your request into tasks → subtasks. Shows you the plan. You confirm once.
Execute Works through each subtask: writes code, runs commands, verifies each step.
Lint Auto-detects your project's linters/formatters and runs them after each task.
Test Auto-detects your test framework, runs relevant tests, catches regressions.
Commit Creates a git commit per task — not one giant commit at the end.
Report Formatted summary: what changed, how to test, expected behavior, decisions made.

Before / After

Without YOLO (5+ prompts)

You: "Add a user preferences feature"
Claude: "Here's a migration..."
You: "Now create the model"
Claude: "Here's the model..."
You: "Run the tests"
Claude: "Tests pass..."
You: "Commit this"
Claude: "Done"
You: "Now the API endpoint..."

With YOLO (1 prompt)

You: "/yolo Add a user preferences feature
      with API endpoints and tests"

[shows plan, you confirm]
[executes everything]
[delivers full report with
 testing flow and commits]

Same result. One prompt. Zero hand-holding.

Install

Claude Code (plugin)

claude plugin marketplace add Cedillin/yolo
claude plugin install yolo@yolo

Cursor

Copy the .cursor/ directory into your project root, or add the rule manually:

# From your project root
curl -sL https://raw.githubusercontent.com/Cedillin/yolo/main/.cursor/rules/yolo.mdc \
  -o .cursor/rules/yolo.mdc --create-dirs

Codex (OpenAI)

# From your project root
curl -sL https://raw.githubusercontent.com/Cedillin/yolo/main/.codex/instructions.md \
  -o .codex/instructions.md --create-dirs

GitHub Copilot

# From your project root
curl -sL https://raw.githubusercontent.com/Cedillin/yolo/main/.github/copilot-instructions.md \
  -o .github/copilot-instructions.md --create-dirs

OpenCode / Other agents

Copy AGENTS.md into your project root:

curl -sL https://raw.githubusercontent.com/Cedillin/yolo/main/AGENTS.md -o AGENTS.md

Any agent (via npx skills)

npx skills add Cedillin/yolo

Usage

Say /yolo followed by what you want:

/yolo Add dark mode toggle to the settings page
/yolo Refactor the auth module to use JWT tokens instead of sessions
/yolo Fix the N+1 query in the dashboard endpoint and add a test

Or use natural language triggers:

Just do it — implement the entire payment integration
Go full auto on the user onboarding flow

What you'll see

  1. Plan — a formatted table of tasks and subtasks
  2. Progress — status updates as each task completes
  3. Report — final summary with:
    • Changelog of everything done
    • Git commits created
    • Step-by-step testing instructions
    • Expected behavior and edge cases
    • Decisions made and notes

Supported Languages

YOLO is language and framework agnostic. It auto-detects your project's tooling:

Linters / Formatters

Language Tools detected
PHP Pint, PHP CS Fixer
JavaScript/TypeScript ESLint, Prettier, Biome
Python Ruff, Black
Ruby RuboCop
Rust rustfmt
Go golangci-lint, gofmt
Elixir mix format
Swift SwiftLint
C/C++ ClangFormat
Deno deno fmt + deno lint
Any Makefile lint/fmt targets

Test Frameworks

Language Frameworks detected
PHP Pest, PHPUnit
JavaScript/TypeScript Jest, Vitest
Python pytest
Rust cargo test
Go go test
Ruby RSpec
Elixir ExUnit
Java/Kotlin Gradle, Maven
Any Makefile test target

Agent Support

Agent Mechanism File
Claude Code Plugin (skill system) skills/yolo/SKILL.md
Cursor Rule file .cursor/rules/yolo.mdc
Codex Instructions .codex/instructions.md
GitHub Copilot Copilot instructions .github/copilot-instructions.md
OpenCode / Others AGENTS.md AGENTS.md

How It Works

┌──────────────────────────────────────────────────────────┐
│                     /yolo <task>                          │
└──────────────┬───────────────────────────────────────────┘
               ▼
┌──────────────────────────┐
│  Phase 1: PLAN           │
│  ├─ Analyze task         │
│  ├─ Break into tasks     │
│  ├─ Define subtasks      │
│  └─ Present plan ────────┼──▶ User confirms
└──────────────┬───────────┘
               ▼
┌──────────────────────────┐
│  Phase 2: EXECUTE        │◀─── Loop per task
│  ├─ Run subtasks         │
│  ├─ Verify each step     │
│  ├─ Auto-lint            │
│  ├─ Run tests            │
│  ├─ Git commit           │
│  └─ Report progress      │
└──────────────┬───────────┘
               ▼
┌──────────────────────────┐
│  Phase 3: REPORT         │
│  ├─ Summary stats        │
│  ├─ Changelog            │
│  ├─ Testing flow         │
│  ├─ Expected output      │
│  └─ Notes & decisions    │
└──────────────────────────┘

License

MIT

About

Full autonomous development mode for AI coding agents. Plan, execute, test, commit, report — one prompt.

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors