GitHub contribution graphs are essentially vanity metrics

I've been implementing this kind of "activity heartbeat" for a few of our internal utility repos to ensure the automation pipelines are firing correctly and the repos stay "warm." Here is a practical tutorial on setting up a forward-looking streak.

My Automated Contribution Workflow

The goal is to simulate realistic human behavior. Pushing exactly one commit at midnight every day looks fake. To make it look authentic, I configured the script to randomize both the volume and the timing of commits.

• Weekday volume: 2 to 7 commits (randomized)
• Weekend volume: 1 to 2 commits (randomized)
• Time window: Randomly distributed between 09:00 and 18:00
• Trigger: Scheduled via cron (UTC) with a manual override option

Implementation Guide

1. Repository Setup

First, create a new repository (e.g., daily-activity). I highly recommend setting this to Private. If you do, make sure to go to your profile settings and enable "Include private contributions" so the green squares actually show up on your public graph. Leave the repo completely empty—no README or .gitignore.

2. The Workflow Configuration

You need to create a workflow file in the .github/workflows/ directory. This YAML file tells GitHub to spin up a Linux runner daily, modify a text file, and commit those changes back to the repo.

name: Forward Daily Contributions

on:
  schedule:
    # Runs every day at 00:00 UTC
    - cron: '0 0 * * *'
  # Allows you to run it manually from the Actions tab
  workflow_dispatch:

jobs:
  forward-commit:
    runs-on: ubuntu-latest
    permissions:
      contents: write

    steps:
      - name: Checkout repository
        uses: actions/checkout@v4

      - name: Generate today's commits
        run: |
          # 🔴 REPLACE WITH YOUR ACTUAL GITHUB EMAIL
          git config user.name "github-actions[bot]"
          git config user.email "[email protected]"

          # Day of week: 1=Monday ... 7=Sunday
          DOW=$(date +%u)

          # Realistic commit counts
          if [ $DOW -ge 6 ]; then
            # Weekend: 1–2 commits
            COMMITS=$((RANDOM % 2 + 1))
          else
            # Weekday: 2–7 commits
            COMMITS=$((RANDOM % 6 + 2))
          fi

          echo "Today is day $DOW. Making $COMMITS commit(s)."

          for i in $(seq 1 $COMMITS); do
            # Create/update a log file
            echo "$(date -u) - Commit #$i" >> log.txt
            git add log.txt

            # Random hour between 9 AM and 6 PM
            HOUR=$((RANDOM % 9 + 9))
            MINUTE=$((RANDOM % 60))
            COMMIT_TIME="$(date +'%Y-%m-%dT')$(printf "%02d:%02d:00" $HOUR $MINUTE)"

            git commit -m "chore: daily update #$i" --date="$COMMIT_TIME"
          done

      - name: Push changes
        run: git push

This setup removes the manual overhead of maintaining a streak. By using the --date flag in the git commit command, the script can backdate the commit to a random time during business hours, making the activity look natural rather than bot-driven.