agent-friendly-action

GitHub Action that checks AI agent readiness on every pull request: it comments with how the PR changes the repo's Agent Friendly Code score — "this PR drops your Claude Code score by 4.1 points because it removed CI config."

What it does

On every pull request, the action:

1. Scores the PR head (the checked-out tree) and the base commit (in a temporary git worktree) with the bundled scorer.
2. Posts one comment with the overall and per-agent change and the checks that moved, and edits that same comment on later pushes.

It reads files only — nothing from the PR is executed — and calls no server besides the GitHub API. The checks and per-agent weights are the same as on the website.

Install

Walkthrough and FAQ: https://agentfriendlycode.com/action.

# .github/workflows/agent-friendly.yml
name: Agent-friendly score diff

on:
  pull_request:
    branches: [main]

permissions:
  contents: read
  pull-requests: write

jobs:
  score-diff:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
        with:
          fetch-depth: 0
      - uses: hsnice16/agent-friendly-action@v0
        with:
          agents-badge-token: ${{ secrets.AGENTS_BADGE_TOKEN }}

AGENTS_BADGE_TOKEN is an opt-in marker, not a credential: set it to any non-empty value in the repo's secrets to enable the comment. Unset, the action exits silently — which is also what happens on fork PRs, where secrets aren't passed. fetch-depth: 0 keeps the base commit local; without it the action fetches it from origin.

Use the pull_request trigger. The action also accepts pull_request_target, but that runs with a write token on fork PRs; it is safe only because the action never executes PR code, so don't add steps that do.

Inputs

Input	Required	Default	Description
agents-badge-token	yes	—	Opt-in marker. The action exits silently when empty.
github-token	no	${{ github.token }}	Posts and edits the comment. A GitHub App installation token or a PAT also works.
base-ref	no	the PR's base commit	A branch, tag, SHA or revision such as HEAD~1 to compare against instead.

Why opt-in?

A template or starter repo can ship this workflow without it commenting on every copy; only repos that set the secret get the comment.

Self-contained

The scorer is vendored from agent-friendly-code and bundled into dist/, so the action keeps working if the website is down.

Related

• agent-friendly-code — the website and the canonical scorer.
• agent-friendly-skill — the same scorer in your editor; scores the local repo and recommends a model class. Install: npx skills add hsnice16/agent-friendly-skill#v0. Details: https://agentfriendlycode.com/skill.

Changelog

CHANGELOG.md.

Sponsor

github.com/sponsors/hsnice16.

License

MIT