Disclaimer: this is one of those "because you can" projects. There are many, many more useful things you can do with GitHub Actions - but if you really, really want to it's flexible enough to even let you play chess!

Play chess in your GitHub repository. Create as many concurrent games as you want and play chess without ever leaving your favorite IDE.

Add the following workflow to one of your repositories, for example as .github/workflows/chess.yaml:

name: Chess!

on:
  push:
    paths:
      - '**.pgn'

jobs:
  play:
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v1

    - uses: cschleiden/chess-action@v1
      with:
        games: '**/*.pgn'

    - name: Commit changed files
      uses: stefanzweifel/git-auto-commit-action@v2.3.0
      with:
        commit_message: Played some chess!
        branch: master
        file_pattern: '*.pgn'
      env:
        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

Then, create a file like game1.pgn anywhere in your repository and enter your first move in Portable Game Notation, for example:

1. e4

Let's commit and push.

$ git add game1.pgn
$ git commit -m "My first move"
$ git push

Now, the action will run, read the game, make its move and commit the result back to the branch. So after a bit of time, let's pull:

$ git pull

And refresh our game1.pgn:

1. e4 e5

{
   +------------------------+
 8 | r  n  b  q  k  b  n  r |
 7 | p  p  p  p  .  p  p  p |
 6 | .  .  .  .  .  .  .  . |
 5 | .  .  .  .  p  .  .  . |
 4 | .  .  .  .  P  .  .  . |
 3 | .  .  .  .  .  .  .  . |
 2 | P  P  P  P  .  P  P  P |
 1 | R  N  B  Q  K  B  N  R |
   +------------------------+
     a  b  c  d  e  f  g  h
}

The action has moved its pawn to e5 and for our convenience added a nice ASCII diagram of the current state of the board.

To add our next move, just modify the file again, commit, push, etc..

Valid inputs to customize are:

Glob pattern for finding games in the repository, defaults to **/*.pgn

The time the chess engine is allowed to think about the next move in milliseconds, defaults to 1 000 (one second).

The way the workflow is defined, it will always read all game files (default **.pgn) and play the next move, so if you have multiple concurrent games, make sure to send your moves in a single push (doesn't have to be single commit).

Also, this pretty much uses some default stockfish settings, I'm not really familiar with UCI engines, so there's probably room for improvement :).

This project makes use of some great libraries which do all the heavy lifting:

• https://github.com/jhlywa/chess.js - To parse, modify, and output PGN and FEN.
• https://github.com/nmrugg/stockfish.js - To actually play.