GitHub Action Template Finder
Pricing
Pay per usage
GitHub Action Template Finder
Find high-quality GitHub Actions workflow templates by category (deployment / testing / security / Claude Code). Each workflow gets a quality score based on parent repo stars, complexity, tool diversity, and freshness. Useful for DevOps teams bootstrapping new repos.
Pricing
Pay per usage
Rating
0.0
(0)
Developer
Yanlong Mu
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
3 days ago
Last modified
Categories
Share
Find production-grade GitHub Actions workflow templates by category — deployment, testing, security scanning, and Claude Code integrations — without manually clicking through 10,000+ public repos. This Actor searches the GitHub code search API, filters by parent-repo stars, fingerprints which tools each workflow uses, then ranks the results by quality. The output is a copy-paste–ready ranked list of .yml files you can drop into your .github/workflows/ directory.
Part of Ian Mu's 100-actor portfolio at github.com/ianymu. Inspired by the post-Stop self-verification flow at github.com/ianymu/claude-verify-before-stop.
What does GitHub Action Template Finder do?
It queries the GitHub code search API for workflow YAML files in the .github/workflows/ directory across all of GitHub, scoped by four supported categories: deployment (Vercel / AWS / GCP / Azure), testing (Playwright / Jest / Cypress), security (CodeQL / Trivy / Snyk / Semgrep), and Claude Code (workflows that use anthropics/claude-code-action). For each hit it fetches the parent repo's star count, downloads the raw workflow file, counts lines, fingerprints which third-party Actions are used, and assigns a 0–100 quality score. Results are sorted high-to-low and pushed to the dataset.
Why use it?
- Bootstrap new repos faster. Find a battle-tested deploy workflow used by a 100k-star project and adapt it, instead of writing from scratch.
- Catalog the Claude Code ecosystem. This is currently the easiest way to see real-world
anthropics/claude-code-actionusage patterns across GitHub. - Audit your own stack. Compare your workflows to top-rated equivalents in your category and find missing pieces (cache, security scanning, etc.).
How to use it
- Open the Input tab and pick your categories (default: all four).
- Set
minRepoStarsto filter out hobby repos (default: 500). - Set
maxResults(default: 30) — higher = slower because of GitHub rate limits. - Important: set a
GITHUB_TOKENenvironment variable (a fine-grained PAT with public_repo scope is enough). The GitHub code search API requires auth. - Click Start and inspect the Output tab when finished.
Input
| Field | Type | Default | Description |
|---|---|---|---|
categoryFilter | array | ["deployment", "testing", "security", "claude-code"] | Categories to fetch |
minRepoStars | int | 500 | Skip workflows whose repo has fewer stars |
maxResults | int | 30 | Stop after this many qualified templates |
Output
Each item in the dataset:
{"repo": "vercel/next.js","workflowPath": ".github/workflows/deploy.yml","workflowName": "Deploy to Production","categories": ["deployment", "vercel"],"lineCount": 87,"toolsUsed": ["vercel-action", "node-setup", "cache"],"rawUrl": "https://raw.githubusercontent.com/vercel/next.js/canary/.github/workflows/deploy.yml","repoStars": 128000,"repoLanguage": "JavaScript","repoPushedAt": "2026-05-19T03:12:44Z","qualityScore": 95}
You can download the dataset in JSON, CSV, HTML, or Excel. A Markdown summary (workflow-templates.md) is also written to the default key-value store.
Data table
| Field | Meaning |
|---|---|
repo | The repository hosting the workflow |
workflowName | Value of name: at the top of the workflow YAML |
categories | Inferred categories (primary + detected from content) |
toolsUsed | Fingerprinted Actions: vercel-action, docker, claude-code, codeql, etc. |
lineCount | Workflow YAML line count |
qualityScore | 0–100, weighted by stars / size / tool diversity / has name / freshness |
Pricing / Cost estimation
Each run hits GitHub's code-search API four times (once per category), plus one repo-meta call and one raw-file fetch per candidate. With default settings (maxResults: 30) and a personal GITHUB_TOKEN, expect ~150 API calls total and well under one compute unit per run on Apify Free.
Tips
- Rate limit: GitHub gives 30 code-search requests/min, 5,000 REST requests/hour for an authenticated user. Bump
maxResultscarefully. - Boost recall: lower
minRepoStarsto find smaller-but-recently-active repos. - Claude Code only: pass
["claude-code"]ascategoryFilterfor an ecosystem map.
FAQ, disclaimers, and support
- Is scraping GitHub legal? This Actor uses GitHub's official REST API. You must comply with GitHub's API terms.
- GitHub code search misses some results. GitHub code search indexes default branches of public repos only. Some workflows on non-default branches will be missed.
- Issues / feature requests: open an issue at github.com/ianymu.
Author: Ian Mu — One-person AI company experiments at github.com/ianymu.