generate-og-image AGENTS.md

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

This is a GitHub Action that generates Open Graph (OG) images from markdown files in pull requests. The action reads frontmatter configuration from .md/.mdx files and creates social media preview images using Puppeteer and a web component.

Development Commands

Build and Development

# Type checking
bun --bun tsc --noEmit

# Build for development
bun build src/main.ts --outdir dist --target node

# Build for release (compiled binary)
bun build src/main.ts --compile --outfile dist/index.js

# Run tests
bun test

# Format code
biome format --write

Code Quality

• Uses Biome for formatting and linting with tab indentation and double quotes
• Pre-commit hooks run biome check --write via lefthook
• Type checking with TypeScript

Architecture

Core Flow (main.ts:17-56)

1. Validates GitHub environment (PR context only)
2. Gets repository configuration via getRepoProps()
3. Finds markdown files in PR with findFile()
4. For each file with OG config:
   • Generates HTML template with generateHtml()
   • Creates image using Puppeteer via generateImage()
   • Commits the image file with commitFile()
   • Posts preview comment (unless disabled)

Key Modules

• find-file.ts: Discovers markdown files in PR, filters by patterns, extracts frontmatter
• generate-image.ts: Uses Puppeteer with Chrome to screenshot HTML into base64 image
• file-filter.ts: Applies glob patterns to ignore specific files (default: /README.md)
• repo-props.ts: Merges GitHub Action inputs with defaults
• github-api.ts: Octokit client for GitHub API operations

Data Flow

• Input: Markdown files with ogImage frontmatter → File discovery & filtering → HTML generation → Image rendering → Git commit & comment

Docker Environment

• Runs in Docker container with Chrome executable at /usr/bin/google-chrome-stable
• Uses action.yml to define GitHub Action interface