skill
ai-skills
A collection of skills for Claude, curser, antigravity, IBM bob or any AI IDE
Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/Vishuvishu/ai-skills itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 26, 2026ai-skills
A collection of skills for Claude or any AI Agents — drop-in instruction sets that make Claude smarter at specific tasks.
Available Skills
| Skill | What it does |
|---|---|
backend-security | Getting scraped, spammed, or hit by bots? Use this skill and ask the Claude or any AI agent to harden your Node.js backend — from IP banning and rate limiting to CAPTCHA and secure headers. For turnstile you also need the changes in the frontend |
seo-optimization | Complete master blueprint for Technical and On-Page SEO, centralized metadata engines, JSON-LD structured data schemas, dynamic XML sitemaps, robots.txt crawler directives, and Core Web Vitals across modern web applications. |
vercel-nextjs | Deploying on Vercel or optimizing Next.js? Use this skill to configure vercel.json for custom regions, cold start reduction, functions, security headers, redirects, crons, and image optimization. |
What is a skill?
A skill is a folder with a SKILL.md file inside. The AI reads it at the start of a conversation and follows its instructions whenever the task is relevant — no manual activation needed.
Installation
Pick the skills you want from the skills/ folder, then follow the steps for your tool.
- Cursor, GitHub Copilot, Antigravity IDE — and any tool supporting
.agents/skills/(Docs: Cursor, Copilot, Antigravity) - Claude.ai (Docs)
- IBM Bob (Docs)
Cursor, GitHub Copilot, Antigravity IDE
These tools all follow the Agent Skills open standard and auto-discover skills from .agents/skills/ in your project root. The steps are the same for all of them.
Project-level (recommended):
Copy the skill folder(s) you want from skills/ into .agents/skills/ in your project:
your-project/
└── .agents/
└── skills/
└── backend-security/ ← paste the whole folder here
├── SKILL.md
└── resources/
└── ...
The agent picks them up automatically when you start a new conversation.
Global (all projects):
Do the same but under ~/.agents/skills/ instead.
Tool-specific notes:
These tools might also support the tools specific folders instead of .agents, but using the .agents is ideal way.
| Tool | Also checks | Global path |
|---|---|---|
| Cursor | .cursor/skills/ | ~/.cursor/skills/ |
| GitHub Copilot | .github/skills/, .claude/skills/ | ~/.copilot/skills/ |
| Antigravity IDE | — | ~/.gemini/antigravity/skills/ |
So if you already have a .cursor/skills/ or .github/skills/ folder, you can put skills there instead — it works the same way.
Claude.ai
Claude.ai natively supports Custom Agent Skills by uploading them as .zip files (available on Pro, Max, Team, and Enterprise plans with code execution enabled).
- Compress your desired skill folder (containing
SKILL.mdand any optionalresources/folder) into a.zipfile. - Open claude.ai and navigate to Settings → Features.
- Locate the Custom Skills section and upload your
.zipfile. - Claude will automatically discover and trigger the skill when relevant to your conversation.
IBM Bob
Note: Skills require Advanced mode in Bob. Make sure it's enabled before starting.
Copy the skill folder(s) you want from skills/ into .bob/skills/ in your project:
your-project/
└── .bob/
└── skills/
└── backend-security/ ← paste the whole folder here
├── SKILL.md
└── resources/
└── ...
Bob will automatically activate the skill when relevant. If not you can also mention liek this in your prompt's start "read the skill-name ( the skills name only ) before proceeding"
Global (all projects): Do the same but under ~/.bob/skills/ instead.
To skip Bob's activation prompt, go to Settings → Auto-Approve → Enable "Always allow skills".
Skill not activating?
If your tool isn't picking up a skill automatically, you can always trigger it manually by referencing the skill file directly in your message:
"Follow the instructions in
.agents/skills/backend-security/SKILL.mdand review my app for security issues."
Or just paste the contents of SKILL.md directly into the chat as context. Every tool supports this as a fallback.