This product was not featured by Product Hunt yet. It will not be visible on their landing page and won't be ranked (cannot win product of the day regardless of upvotes).
Product upvotes vs the next 3
Waiting for data. Loading
Product comments vs the next 3
Waiting for data. Loading
Product upvote speed vs the next 3
Waiting for data. Loading
Product upvotes and comments
Waiting for data. Loading
Product vs the next 3
Loading
skillgrab
One command to install AI skills that match your project
skillgrab scans your project, detects your stack — Next.js, Django, Go, Flutter, 60+ frameworks — and installs matching AI agent skills from skills.sh in one command. Works with Claude Code, Cursor, Cline, Codex, and 20+ agents. Includes an MCP server so your AI agent can call it directly from chat. Status and update commands to manage installed skills. Trusted-owner ranking and GitHub validation for security. npx skillgrab
I started using AI coding agents heavily this year << Claude Code mostly >> and I kept hitting the same friction: every new project, I'd spend 20-30 minutes browsing skills.sh (which has 90,000+ skills), picking the ones that fit my stack, running npx skills add one by one, and usually missing half the relevant ones. Or I'd give up and just have a generic setup that didn't really help my specific project.
So I built skillgrab. One command:
npx skillgrab
What it does
Scans your project files, package.json, requirements.txt, pubspec.yaml, go.mod, Dockerfile, vercel.json, even your README.
Detects your stack. Not just "JavaScript", specifically Next.js + Supabase + Stripe + Clerk, or Django + Celery + SQLAlchemy, or Flutter + Firebase, etc.
Queries skills.sh's API live, ranks results by trusted owners (Anthropic, Vercel, Supabase, Stripe, Clerk, OpenAI, Microsoft, Google, …) and install count.
Dedupes by skill name so you don't get two copywriting skills overwriting each other in ~/.claude/skills/.
Validates each candidate against GitHub before showing it to you, because skills.sh's search sometimes returns slugs that don't actually exist in the repo (fuzzy match artifacts).
Shows you an interactive checklist, uncheck anything you don't want.
Installs via npx skills add, grouped per repo so each repo is cloned only once even if you're installing 5 skills from it.
It also reads your README looking for non-code hints ("landing page", "pricing", "waitlist", "SEO", "outreach") and asks whether you want marketing / design / sales skills included. Opt-in, never forced.
MCP server (new in v0.6)
skillgrab also works as an MCP server… add it to Claude Desktop or Cursor and your agent can call skillgrab_recommend directly from chat. No terminal needed. Three tools: recommend, install, status.
Auto-detects 20+ agents (Claude Code, Cursor, Cline, Codex, Continue, Gemini CLI, Warp, Windsurf, GitHub Copilot, Roo, and more) and installs to all of them at once.
npx skillgrab status - see what's installed, cross-referenced against the live registry.
npx skillgrab update - refresh all installed skills in one command.
--only-trusted - restrict to verified owners only (Anthropic, Vercel, Supabase, Stripe, etc.).
What I learned building it
Honestly more than I expected:
skills.sh is a Next.js SPA and their search endpoint isn't documented… I had to reverse-engineer /api/search from the rendered bundle.
The skills CLI accepts --skill <name> but not --skill a,b,c (it treats the whole comma-separated string as one literal name). Burned me for an hour before I realized.
Some repos store skills at the root, others under <plugin-name>/skills/<id>/SKILL.md - the validator handles both by doing a recursive GitHub tree lookup.
npx skills add owner/repo/skill fails - the correct invocation is npx skills add owner/repo --skill skill.
Every gotcha is fixed in the code. v0.6.1 is the current release.
PRs for new detectors are welcome and take ~50 lines each.
What's next
Watch mode: re-scan when package.json changes.
Pre-built skill packs for common project archetypes (SaaS starter, ML pipeline, etc.).
Community skill pack registry - let anyone publish curated packs.
Try it
npx skillgrab --dry-run
--dry-run shows the plan without installing. Once you like it, drop the flag.
Open source (MIT), written in TypeScript. Would love feedback - especially on detection heuristics for stacks I don't have fixtures for yet (Elixir, Clojure, .NET).
Thanks for the upvotes and comments - happy to answer anything.
“One command to install AI skills that match your project”
skillgrab was submitted on Product Hunt and earned 4 upvotes and 1 comments, placing #67 on the daily leaderboard. skillgrab scans your project, detects your stack — Next.js, Django, Go, Flutter, 60+ frameworks — and installs matching AI agent skills from skills.sh in one command. Works with Claude Code, Cursor, Cline, Codex, and 20+ agents. Includes an MCP server so your AI agent can call it directly from chat. Status and update commands to manage installed skills. Trusted-owner ranking and GitHub validation for security. npx skillgrab
On the analytics side, skillgrab competes within Open Source, Developer Tools and GitHub — topics that collectively have 621.3k followers on Product Hunt. The dashboard above tracks how skillgrab performed against the three products that launched closest to it on the same day.
Who hunted skillgrab?
skillgrab was hunted by Ismael Briasco. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
For a complete overview of skillgrab including community comment highlights and product details, visit the product overview.
Ismael Briasco
Hey Product Hunt 👋
I'm Ismael. Thanks for checking out skillgrab.
The problem that kept bugging me
I started using AI coding agents heavily this year << Claude Code mostly >> and I kept hitting the same friction: every new project, I'd spend 20-30 minutes browsing skills.sh (which has 90,000+ skills), picking the ones that fit my stack, running npx skills add one by one, and usually missing half the relevant ones. Or I'd give up and just have a generic setup that didn't really help my specific project.
So I built skillgrab. One command:
What it does
Scans your project files, package.json, requirements.txt, pubspec.yaml, go.mod, Dockerfile, vercel.json, even your README.
Detects your stack. Not just "JavaScript", specifically Next.js + Supabase + Stripe + Clerk, or Django + Celery + SQLAlchemy, or Flutter + Firebase, etc.
Queries skills.sh's API live, ranks results by trusted owners (Anthropic, Vercel, Supabase, Stripe, Clerk, OpenAI, Microsoft, Google, …) and install count.
Dedupes by skill name so you don't get two copywriting skills overwriting each other in ~/.claude/skills/.
Validates each candidate against GitHub before showing it to you, because skills.sh's search sometimes returns slugs that don't actually exist in the repo (fuzzy match artifacts).
Shows you an interactive checklist, uncheck anything you don't want.
Installs via npx skills add, grouped per repo so each repo is cloned only once even if you're installing 5 skills from it.
It also reads your README looking for non-code hints ("landing page", "pricing", "waitlist", "SEO", "outreach") and asks whether you want marketing / design / sales skills included. Opt-in, never forced.
MCP server (new in v0.6)
skillgrab also works as an MCP server… add it to Claude Desktop or Cursor and your agent can call skillgrab_recommend directly from chat. No terminal needed. Three tools: recommend, install, status.
Multi-agent + status management
Auto-detects 20+ agents (Claude Code, Cursor, Cline, Codex, Continue, Gemini CLI, Warp, Windsurf, GitHub Copilot, Roo, and more) and installs to all of them at once.
npx skillgrab status - see what's installed, cross-referenced against the live registry.
npx skillgrab update - refresh all installed skills in one command.
--only-trusted - restrict to verified owners only (Anthropic, Vercel, Supabase, Stripe, etc.).
What I learned building it
Honestly more than I expected:
skills.sh is a Next.js SPA and their search endpoint isn't documented… I had to reverse-engineer /api/search from the rendered bundle.
The skills CLI accepts --skill <name> but not --skill a,b,c (it treats the whole comma-separated string as one literal name). Burned me for an hour before I realized.
Some repos store skills at the root, others under <plugin-name>/skills/<id>/SKILL.md - the validator handles both by doing a recursive GitHub tree lookup.
npx skills add owner/repo/skill fails - the correct invocation is npx skills add owner/repo --skill skill.
Every gotcha is fixed in the code. v0.6.1 is the current release.
Supported stacks
React, Next.js, Vue, Nuxt, Angular, Svelte, SvelteKit, Astro, Remix, Solid, Tailwind, Chakra, MUI, Express, Fastify, Hono, tRPC, Prisma, Drizzle, Supabase, Firebase, Clerk, NextAuth, Stripe, OpenAI, Anthropic, Vercel AI SDK, LangChain, React Native, Expo, Electron, Vitest, Jest, Playwright, Cypress, Django, Flask, FastAPI, PyTorch, TensorFlow, Celery, SQLAlchemy, Flutter, Swift, Android, Go, Rust, Rails, Laravel, Phoenix, Docker, Vercel, Netlify, Fly.io, Cloudflare Workers, Terraform, GitHub Actions.
PRs for new detectors are welcome and take ~50 lines each.
What's next
Watch mode: re-scan when package.json changes.
Pre-built skill packs for common project archetypes (SaaS starter, ML pipeline, etc.).
Community skill pack registry - let anyone publish curated packs.
Try it
--dry-run shows the plan without installing. Once you like it, drop the flag.
Open source (MIT), written in TypeScript. Would love feedback - especially on detection heuristics for stacks I don't have fixtures for yet (Elixir, Clojure, .NET).
Thanks for the upvotes and comments - happy to answer anything.
Ismael (@briascoi)