Research

Give Codex or Claude Real-Time Web Search Through a Grok Endpoint

A 181-star skill that routes your agent's lookups — version numbers, API docs, error messages — through your own OpenAI-compatible Grok endpoint and returns JSON with sources.

Learn Agent Skills by Example with DeepLearning.AI's Course Files

The official artifacts repo for the Agent Skills with Anthropic short course ships the course's prompts, notes, and data files — clone it and study the same task with and without skills.

claude-osint: 8 Skills That Turn Claude Into an Authorized Recon Analyst

A 2,583-star library of eight SKILL.md files that give Claude external-recon methodology for assets you own or are authorized to assess, with hard boundaries against active exploitation.

agent-skills-with-anthropic: Andrew Ng's Agent Skills Course, In Chinese

Datawhale's 1,502-star study companion translates Andrew Ng's DeepLearning.AI short course on Agent Skills into Chinese, adds per-lesson knowledge notes and code walkthroughs, and pairs every lesson with a named reviewer — the fastest on-ramp for Chinese-speaking teammates.

Apify Agent Skills: Describe the Data, Let Your Agent Do the Scraping

Apify's official 2,365-star skill pack gives coding agents Actor selection across 130-plus curated scrapers with fallback to the 30,000-strong Apify Store — five skills that cover scraping, Actor building, and SDK integration.

auto-empirical-research-skills: 1,096 Econometrics Skills Backed by Benchmarks

A 3.6k-star Stanford REAP × CoPaper.AI collection that packages empirical-research methods — from data cleaning to journal submission — as agent skills, with 19 benchmark tasks and a public scorer that checks your agent actually runs them correctly.

browserbase/skills: Give Claude Code a Full Browser Automation Toolkit

A 3.7k-star official skill collection that turns Claude Code into a working browser agent — remote sessions with CAPTCHA solving, DevTools trace capture, adversarial UI testing, and one-command QA on your localhost dev server.

Build Production-Grade AI Agents with Context Engineering Skills

17 skills focused on context engineering and harness engineering principles — covering context fundamentals, multi-agent patterns, memory systems, tool design, evaluation, and self-improvement loops for production AI agents.

Programmatic Access to NotebookLM with notebooklm-py

An unofficial Python API, CLI, MCP server, and agent skill for Google NotebookLM (now Gemini Notebook) — generate audio overviews, video explainers, slide decks, infographics, quizzes, and reports, and use it as cross-session memory for Claude Code.

Auto-Generate Agent Skills with Microsoft's SkillOpt

SkillOpt trains reusable best_skill.md artifacts for frozen LLM agents through trajectory-driven edits — measured gains of +23.5 accuracy points on SearchQA across GPT-5.5, Codex CLI, and Claude Code CLI without fine-tuning the model.

Turn Any AI Agent Into a Scientist with Scientific Agent Skills

A collection of 161 ready-to-use skills that let your AI agent query PubChem, RDKit, AlphaFold, and other scientific databases and Python packages, turning a research pipeline into a single conversation.

Master Context Engineering: Build Better AI Agent Systems

How to use muratcankoylan/agent-skills-for-context-engineering to design, optimize, and evaluate context windows for production AI agents.

Research Any Topic in Minutes with the Last 30 Days Skill

A hands-on guide to using mvanhorn/last30days-skill to let your AI agent research topics across Reddit, X, YouTube, HN, and more.