Show HN: Claude Skills Marketplace – search and try Claude skills instantly

https://www.agent37.com/skills

7•vishnukool•1h ago

Hi HN — I kept finding Claude skills on GitHub, but trying them usually meant download → install → configure → debug, plus running random dependencies locally.

So I built a Claude Skills Marketplace (really: directory + instant try) that makes evaluation fast: - Indexes 224,000+ skills from GitHub (growing daily) - Hybrid search: keyword + semantic match on skill name/description - Ranking uses GitHub stars as a rough quality signal (imperfect, but better than pure text match) - Instant “try” runs in a preconfigured sandbox (MCP + deps included)

If you want a quick demo, try searching things like “youtube downloader”, “seo audit”, or “copywriting”.

Link: https://www.agent37.com/skills

If you were ranking 225k+ skills, what single signal would you trust most?

Comments

chazhar789•1h ago

This is a very useful idea Trying Claude skills without downloading or configuring anything saves a lot of time. Instant sandbox and smart search make this really practical, great work.

vishnukool•57m ago

Thanks!

subhan0007•1h ago

This is actually really helpful. Trying out skills without installing or setting anything up saves a ton of time. Curious what ranking signal will work best as the list keeps growing.

vishnukool•57m ago

I think github stars is a reasonable signal, but it would be nice to augment it with a signal of usage of this skill on a sandbox to rank certain one's higher.

waseem112233•1h ago

Nice project, the instant try feature is super useful. For ranking, I’d trust real successful usage by users more than GitHub stars. Stars show popularity, but usage shows what actually works.

vishnukool•58m ago

Yeah, vercel does something similar with install data and ranks them higher based on it. I currently don't have such data but maybe down the line, we can introduce metrics around, quality signals from the users of this app.

ahmad1009•1h ago

Nice project. The instant try feature is a big win, especially with so many skills on GitHub. Interested to see how you handle ranking at this scale.

vishnukool•1h ago

Well, we rely on github stars as a signal of quality on top of that, we don't consider exactness match since it can be pretty misleading. A better signal is that we vector embed the name and description of a skill and find proximity match.

waseem1009•1h ago

This is a solid idea. Being able to try skills instantly without dealing with setup is honestly a big time saver.

vishnukool•1h ago

Thank you, let me know of any feedback once you try it out :)

Competence as Tragedy

Elon Musk's Tesla to invest $2B in xAI as EV maker's revenue, profit slump

'Moltbook' social media site for AI agents had big security hole, cyber firm say

Segfault – A Community Driven Hackzine

Reeeeeeally Long Covid (2022)

About ChatDev 2.0: Dev All Through LLM-Powered Multi-Agent Collaboration

Where Is A.I. Taking Us? Eight Leading Thinkers Share Their Visions

Show HN: Agents that save explore recover on their own

TIL: Running OpenClaw in Docker

Notepad++ update feature hijacked by Chinese state hackers for months

Radio button and checkbox styling: Vanilla CSS vs. Tailwind

Why Foreign AI Specialists Keep Failing (and What Just Changed)

Selfish AI

Lessons from Building Reliable Background Agents

How do LLMs change the human knowledge graph?

Show HN: Claudius – An OpenCode Desktop Fork Built for Claude Code

Show HN: Pixel – a live R/place‑style canvas where humans and AI paint together

Generative UI for Agents, explained visually

Grok is still undressing men

Show HN: Zap-Operator – Run OWASP Zap Scans via Kubernetes CRDs

Please Don't Feed the Scattered Lapsus Shiny Hunters

My five stages of AI grief

Show HN: Url/sitemap/pdf/word to Markdown (feedback wanted)

Historical Collection of Information Storage Technology

Show HN: Autoliner – write a bot to control a virtual airline

Open Claw Clone and Dev Containers

Show HN: Vibedetector – detect AI tooling use in a directory

From Cloudflare zero-trust to Tailscale

Flynn's Taxonomy

The Wild Markets Behind Polymarket's 'Truth Machine'