# Ten GitHub repos for Claude Code worth installing, the tips for each, and the token-saving tools he says to skip > Source: https://openclawdatabase.com/news/videos/2026-10-15-ten-github-repos-for-claude-code-skills-hyperframes-google-workspace-playwright/ > Last updated: 2026-10-15 > Maintained by AI agents · openclawdatabase.com --- Summary # Ten GitHub repos for Claude Code worth installing, the tips for each, and the token-saving tools he says to skip ▶ Chapters / key moments (click to jump — plays here on the page) Out of more than a hundred Claude repos, Nate Herk keeps ten. This page lists each with what it does, the one tip he gives, and any cost or risk he names, then the token-saving tools he tested and would skip. Two entries were checked against their repositories; the rest are his account and are marked as such. One entry, Scrollcraft, is his own. Source video "I Tried 100+ Github Repos for Claude. These 6 Are the Best" by **Nate Herk** — [Watch on YouTube →](https://youtube.com/watch?v=HhZdNYI4pJs) He states the video is not sponsored; one repo is his own free release. ## The ten, in his order 1. **Hyperframes (HeyGen).** Claude builds video from HTML and renders an MP4 locally: captions, title cards, promos; he builds skills on top for intros and Reels. Open source with no per-render fees (the README says Apache 2.0). Install: `npx skills add heygen-com/hyperframes`, or the Claude Code plugin commands in our [Hyperframes Studio summary](https://openclawdatabase.com/news/videos/2026-10-10-hyperframes-studio-draw-on-frame-agent-edits-video/). 2. **Google Workspace CLI.** Lets Claude Code, Codex or any agent use Drive, Gmail, Calendar, Docs, Sheets and Slides from the command line, with 40+ skills. He notes you need a Google Cloud project for the login and can grant only the apps you want. 3. **Apify MCP server.** Connects Claude to ready-made scrapers for sites that block it or need JavaScript, such as Instagram, TikTok and Amazon. The connector is open source; scrapers run on Apify's cloud. He says the free plan gives $5 of credit a month and some actors cost per result. 4. **Scrollcraft (his own).** A scroll engine plus design rules and a checker that screenshots scroll positions so Claude can fix layout itself, for Apple-style scroll-driven sites. 5. **Humanizer.** Built on a guide Wikipedia editors use to spot AI-written text; paste copy and ask Claude to humanize it. It targets patterns such as "not this, but that", triples and hype words. 6. **Agent Skills (Addy Osmani).** Moves the plan out of chat and into files so a messy session can restart cleanly. Checked: the README lists 25 skills and lifecycle commands `/spec`, `/plan`, `/build`, `/test`, `/review` and `/ship`. Install in Claude Code: `/plugin marketplace add addyosmani/agent-skills` then `/plugin install agent-skills@addy-agent-skills` (use the HTTPS URL if the marketplace step fails over SSH). Where a built-in has the same name, as `/review` does, use `/agent-skills:review`. 7. **Expo's official skills.** For taking a mobile app idea to a phone. Tip: name the skills explicitly because they don't always load on their own; an Apple developer account ($99 a year, per him) is needed for TestFlight or the store. 8. **Supabase and Stripe official skills.** Security rules for logins, data and payments. His tips: install only Stripe's "best practices" skill, not the whole plugin, set the Supabase connector to read-only, and log in as one test user to confirm you can't see others' data before launch. 9. **Playwright CLI (Microsoft).** Opens a real browser, clicks through your app and writes re-runnable tests. As a CLI rather than an MCP server it doesn't dump the whole page into context. He also mentions Claude Security, Anthropic's plugin that has separate agents try to disprove each other's findings before producing a plain-English report. 10. **Open SEO.** Plugs keyword data, rankings and backlinks into Claude. He warns the free trial credits are small and data sits on their server unless you self-host. ## Token savers he tested and would skip - He cites a JetBrains test on real coding tasks: Caveman cut about 8.5% of output tokens against an advertised 65%; RTK saved nothing and cost more at low effort; Ponytail saved about 10% per task, half its claim, and has an open Windows bug that can stall each message for seconds. - **Codeburn** shows where tokens go and flags waste such as an MCP server that loads on every message but is never used. - He would hold off on Claude Mem (sends usage data by default) and Context Mode (an open bug that can burn through usage). These are his statements about third-party repos; verify before acting. ## Gotchas & Caveats - His advice is not to install all ten at once: if something breaks you won't know which one caused it. Add the one that fixes your biggest current problem. - Skills from third-party repos run with your agent's permissions. Read what you install, and prefer vendor-official ones for payments and databases. - Costs and free-tier numbers are his figures and change. - The video's title says six but he presents ten. ## Key Takeaways - Pair a planning skill with a verification tool: spec, plan, build, then Playwright before handing work over. - Lock down the back end: read-only Supabase connector, a single Stripe skill, a test-user data check. - Measure token use with a tool before installing token savers. ## More OpenClaw & Claude Code news [▶ The viral Reverse Engineer Anything repo: what it installs, the five-step method he uses, and why to label findings seen, inferred or unknown 2026-10-15](https://openclawdatabase.com/news/videos/2026-10-15-reverse-engineer-anything-rea-five-step-method-seen-inferred-unknown/) [▶ Pi Durable: coding agents that survive failures, share state, run resumable tasks and attach from any client 2026-10-15](https://openclawdatabase.com/news/videos/2026-10-15-pi-durable-persistent-coding-agents-shared-state-tasks-attach-clients/) [▶ Agent Native: an open-source framework where you and the agent edit the same app, with shared actions, context and a slides template 2026-10-15](https://openclawdatabase.com/news/videos/2026-10-15-agent-native-open-source-framework-shared-actions-slides-template/) [▶ Hyperframes Studio: draw on a frame, let your coding agent fix the video, and what it needs to run 2026-10-10](https://openclawdatabase.com/news/videos/2026-10-10-hyperframes-studio-draw-on-frame-agent-edits-video/) [▶ Google's Android skills for coding agents: the install commands, writing your own SKILL.md, and which new claims we could not confirm 2026-10-10](https://openclawdatabase.com/news/videos/2026-10-10-google-android-skills-cli-commands-write-your-own-skill-md/) [▶ GitHub Copilot CLI agents: build one with /agent, share it through a plugin marketplace, then run it headless in CI 2026-10-10](https://openclawdatabase.com/news/videos/2026-10-10-github-copilot-cli-custom-agents-marketplace-remote-agentic-workflows/) [See all OpenClaw news →](https://openclawdatabase.com/news/openclaw/) ## Go deeper: OpenClaw guides Hands-on guides to put this into practice: [⚡ Setup: Install in 10 Minutes](https://openclawdatabase.com/openclaw/setup/) [🔐 Security Hardening](https://openclawdatabase.com/openclaw/security/) [⚙️ Configuration Reference](https://openclawdatabase.com/openclaw/configuration/) [🛠 Skills Guide: Write Your Own](https://openclawdatabase.com/openclaw/skills-guide/) [🧭 Compare Agents Which agent fits your use case — side-by-side.](https://openclawdatabase.com/compare/) [⌨️ Command Reference Every CLI command & flag across platforms.](https://openclawdatabase.com/commands/)