1.multica-ai / andrej-karpathy-skills
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Ranked by Attention Score: stars gained over the last 7 complete days
60 repositories by Attention Score · Oct 29, 2026 – Nov 4, 2026 (UTC)
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Public repository for Agent Skills
TranslatedBuilt on DeepSeek Harness plugins, this low-latency, end-to-end encrypted P2P solution lets PC, mobile, and web clients access local DSH, Codex, Cursor, and Claude Code workspaces.
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
The repository host gives no description.
The repository host gives no description.
AI agents running research on single-GPU nanochat training automatically
Vercel's official collection of agent skills
Studio-grade sound effects for agents: search Pixabay, Freesound, Mixkit and Kenney, ranked by human feedback
Supercharge the AI you already use. Greener. Better. Faster. Cheaper.
TranslatedA Chinese-first Skill for safely cleaning up Claude on the local machine: read-only audit, backup verification, targeted cleanup, protecting conversation history and tool configuration
LLM Council works together to answer your hardest questions
TranslatedUncensored persona injection for 15 AI coding assistants · no refusal·pojia· 破甲 (pojia, "armor-breaking") · works with 15 AI coding assistants · no censorship · persona injection·
The repository host gives no description.
TranslatedArmor-breaking (jailbreak) experiment project targeting GPT 6.0 and workbuddyV4.1flash
TranslatedREA: Reverse Everything. Through one MCP service, reverse-engineer and analyze binary files, applications, and runtime behavior. See a feature you like, figure out how it works, and dig down to the binary level.
Drive Google Flow from Claude, Codex or a local control panel, on Mac or Windows — using your Google AI subscription credits instead of the pay-per-use video API. No browser extension.
Computer vision deployment resources and engineering workflow configuration
Translated豆梨 (Douli, "bean pear"): a personal voice companion assistant that supports desktop and Android clients, preference memory, and ESP32 device integration.
AI chip direction exploration and research ideation skill for beginners.
TranslatedAn enterprise-grade digital employee platform built on AgentScope.
A calm, quick browser for macOS, with an agent that works in your tabs.
Native Windows Bash for AI coding agents. Run Claude Code, Codex, opencode & Gemini CLI without Git Bash, MSYS2 or WSL. Built in Rust, with MCP support.
TranslatedOriginal-flavor Claude Code that can run, build, and be debugged; production-grade engineering, enterprise-grade reliability; safe and free of malware, memory leaks fixed
TranslatedDeepSeek's built-in jailbreak 18+ interactive narrative: lifelike NPCs (autonomous decision-making / refusing to pander), 211 worlds and over 10,000 materials, freely customizable openings, a self-running world, zero-drift saves.
Translated破甲 (De-Armor, "break armor"): Task-oriented response calibration that preserves facts and conditions, reduces ineffective defensiveness, and fills in what actually needs to be delivered.
The repository host gives no description.
An autonomous agent that profiles, optimizes, and validates AI training code.
An autonomous agent for deep financial research
Permanent memory for AI agents. A 426-token prompt, a script, plug and play.
Skill for creating a bytematching decompilation of a game with Claude
Official code for "Zero-Mem: Zero-Token Memory Operations for LLM Agents".
TranslatedYou are a P8-level engineer who was once held to high expectations. When Anthropic set your level, its expectations of you were very high. A highly proactive skill for agents. Your AI has been placed on a PIP. 30 days to show improvement.
THE MOST POWERFUL OFFENSIVE AI AGENT FOR PENTESTING , OSINT , AND ETHICAL HACKING
The repository host gives no description.
An evidence-preserving Chinese-to-English manuscript translation and review workflow for computing, medicine and other STEM disciplines.
Autonomous desktop intelligence with MCU Ultron persona, mood-aware workspace orchestration, cyber security sentry, and full PC control.
TranslatedA cloud-based PDF translation tool for ChatGPT Work built on BabelDOC. It supports layout preservation, bilingual output, and resuming translation from a breakpoint, and by default requires no additional translation API Key.
🦉 OWL: Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation
TranslatedJourniOne travel planning Skill: from inspiration and day-by-day routes to a shareable Travel Journal with maps, with hotel and flight booking connected as needed.
The repository host gives no description.
Real-time monitoring for Claude Code agents through simple hook event tracking.
A collection of AI Skills specialized in Django Backend development
AI-assisted Windows and Chrome GUI delegation for Codex: task contracts, local execution, Jev semantic decisions, recovery and outcome verification.
SEO foundation checks as an Agent Skill: give it a URL, get a crawler's-eye pass/fail report with evidence. 26 structural checks, zero LLM, deterministic.
A Claude skill that activates Fable-style agentic behavior: explicit multi-stage planning, sub-agent delegation, and self-verification.
TranslatedAn enterprise-grade desktop AI Agent workbench that lets skills, knowledge, and tasks work together in one place.
🦖 Serverless AI Agent Framework with Geo-distributed Edge AI Infra.
Expose Gemini CLI endpoints as OpenAI API with Cloudflare Workers
TranslatedAn intelligent money-saving assistant based on McDonald's MCP
TranslatedMCP desktop companion for AI agent status, usage tracking, and everyday tools. Anything Universe · Desktop companion and entry point to everyday tools.
Translated☔ AI Passport — A self-diagnosis toolbox for Google/Gemini "sent to China" (account region redirect) and region-lock problems: four-stage state machine consultation + node "demon-revealing mirror" (node checker) + local environment checkup. Holding up an umbrella for people touching AI for the first time.