this-little-wiggy

"Mayor Quimby gives Ralph the instructions. Ralph executes them."

Type vibes, ship autonomy. Detects complex implementation tasks and wraps them in ralph-loop format with project-specific completion criteria.

Why?

You've already curated your workflows - slash commands, TDD prompts, issue references. Why write more prompts just to make them ralph-loop compatible?

this-little-wiggy lets you keep using your established workflows. Your existing prompts automatically get wrapped in ralph-loop format when they're complex enough to benefit from autonomous execution.

What It Does

Intercepts complex implementation tasks and wraps them with project-specific completion criteria for autonomous execution.

Before: Manually structure every prompt with ralph-loop syntax and completion markers

After: Use your existing slash commands and prompts - this-little-wiggy handles the wrapping

Quick Example

# You type your natural prompt:
claude "build a REST API for todos"

# Or use slash commands with a prefix:
claude "Execute /tdd phase1"

# this-little-wiggy wraps it for ralph-loop:
/ralph-loop:ralph-loop --max-iterations 10 --completion-promise "COMPLETE" "build a REST API for todos

When complete:
- Tests pass
- Lint passes
- Build succeeds

Output <promise>COMPLETE</promise> when done."

Installation

Requirements:

• Claude Code 2.0.22+
• ralph-loop plugin
• Python 3.10+ with pyyaml installed (pip install pyyaml)

Windows users: Ensure Python is in your PATH and pyyaml is installed in the Python environment that Claude Code uses.

Option 1: Via Marketplace (Recommended)

1. Add the marketplace:

claude plugin marketplace add severity1/severity1-marketplace

2. Install the plugin:

claude plugin install this-little-wiggy@severity1-marketplace

3. Restart Claude Code

Verify installation with /plugin command. You should see the this-little-wiggy plugin listed.

Option 2: Local Plugin Installation (for Development)

1. Clone the repository:

git clone https://github.com/severity1/this-little-wiggy.git
cd this-little-wiggy

2. Add the local marketplace:

claude plugin marketplace add /absolute/path/to/this-little-wiggy/.dev-marketplace/.claude-plugin/marketplace.json

Replace /absolute/path/to/ with the actual path where you cloned the repository.

3. Install the plugin:

claude plugin install this-little-wiggy@local-dev

4. Restart Claude Code

Verify installation with /plugin command. You should see "1 plugin available, 1 already installed".

Setup

/this-little-wiggy:init

The wizard:

1. Explores your codebase to discover completion criteria
2. Asks for max iterations (budget-based: 5/10/20/30)
3. Lets you validate the discovered criteria
4. Saves config to .claude/this-little-wiggy/config.yml

Usage

Normal prompts - complex tasks auto-detected:

claude "implement user authentication"
claude "build a REST API for todos"

Slash commands - prefix with a word to prevent premature expansion:

claude "Execute /tdd phase1"
claude "Implement /tdd GH issue 32"

Why the prefix? Without it, slash commands expand immediately before the hook can evaluate them. Adding "Execute", "Implement", or any word first lets the hook wrap the command with ralph-loop.

Bypass:

claude "* just fix this one line"    # * prefix bypasses
claude "/ralph-loop:ralph-loop ..."             # already wrapped

Config

# .claude/this-little-wiggy/config.yml
defaultMaxIterations: 10

ralphWrapper: |
  /ralph-loop:ralph-loop --max-iterations 10 --completion-promise "COMPLETE" "{task}

  When complete:
  - Tests pass
  - Lint passes
  - Build succeeds

  Output <promise>COMPLETE</promise> when done."

Architecture

User prompt → Hook (prompt-evaluator.py) → RALPH-LOOP DIRECTIVE → Claude executes /ralph-loop:ralph-loop

The hook reads your project config and wraps prompts with completion criteria. Claude then evaluates whether to execute the wrapped prompt based on task complexity.

Testing

uv run pytest tests/ -v

License

MIT