WAILL-E

             .--------.       .--------.
            /  .----.  \     /  .----.  \
           |  /  (o)   |     |   (o)  \  |
           |  \        |     |        /  |
            \  `------' \___/ `------'  /
             `--------'-'   `-'---------'
                         ||
                      ___||___
                  _.-'--------'-._
                 /|  ________  |  \
          ((    / | |WAILL-E | |  |    ))
          ))   |  | |________| |  |   ((
         ((  _ |  |  _      _  |  | _  ))
          \_/ \|  | (_)    (_) |  |/ \_/
              |  |            |  /
               \  \__________/  /
                `--------------'
                  /==========\
                 /  O  O  O   \
                /______________\

A minimal, robust framework for fast, tracked, and low-overhead agentic development. Optimized for developers with limited time who need to pick up where they left off without losing context.

Philosophy

1. Short, Tracked Iterations: Avoid context decay by keeping active task state in .waill-e/ACTIVE_SLICE.md.
2. Strict Slicing: Work in vertical slices of functionality.
3. Rigid TDD: Follow the Red-Green-Refactor-Document loop.
4. Terse Communication: Use Caveman style to minimize token overhead and keep responses direct.
5. Interactive Engagement: Use questions/choices dynamically to resolve ambiguities.

Getting Started

To use WAILL-E in a project:

Option A: Run the installer script

curl -fsSL https://raw.githubusercontent.com/devmunoz/WAILL-E/main/install.sh | bash

Specify target path if you want:

curl -fsSL https://raw.githubusercontent.com/devmunoz/WAILL-E/main/install.sh | bash -s -- /path/to/your/project

Option B: Manual setup

1. Copy the framework rules and .agents/ folder into your project root:

   # 0 clone this repo:
   git clone https://github.com/devmunoz/waill-e.git && cd waill-e
   
   # 1.a If your project doesn't have an AGENTS.md, copy as AGENTS.md:
   cp WAILL-E.md /path/to/your/project/AGENTS.md
   
   # 1.b If your project already has an AGENTS.md, copy as WAILL-E.md and reference it:
   cp WAILL-E.md /path/to/your/project/WAILL-E.md
   # Then reference it in your existing AGENTS.md (e.g., "See [WAILL-E.md](WAILL-E.md)").
   
   # 2. Copy the skills folder:
   cp -r /path/to/waill-e/.agents /path/to/your/project/

2. Start iterating: When the agent is first run, it will automatically create the .waill-e folder and initialize .waill-e/ACTIVE_SLICE.md.

Release packaging (maintainers)

The installer downloads this fixed release asset from the latest GitHub release:

• waill-e-package.tar.gz

Build it with:

./scripts/build-release-package.sh

This repository also includes a GitHub Action (.github/workflows/release-package.yml) that builds and uploads dist/waill-e-package.tar.gz automatically when a new tag v* (e.g. vX.Y.Z) is pushed.

Updating Skills (maintainers)

To update the skills (caveman, codebase-design, grill-me, grilling, tdd) to their latest versions from upstream sources, run:

./scripts/update-skills.sh

You can also specify particular skills to update:

./scripts/update-skills.sh tdd grill-me

Acknowledgements

Skills are sourced from mattpocock/skills and JuliusBrussee/caveman, which inspired this framework.