# andrej-karpathy-skills | multica-ai | Knowledge Base | Instagit A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls. GitHub Stars: 59.1k Repository: https://github.com/multica-ai/andrej-karpathy-skills --- ## Articles ### [Skills vs CLAUDE.md vs Cursor Rules for Effectiveness: A Complete Technical Guide](/multica-ai/andrej-karpathy-skills/skills-vs-claude-md-vs-cursor-rules) Unlock AI assistant effectiveness with our technical guide on Skills CLAUDE.md and Cursor rules. Learn how this layered governance model boosts reliability and reduces cognitive load. - Tags: deep-dive - Published: 2026-04-19 ### [How to Create a Feedback Loop for Guideline Improvement in LLM-Assisted Development](/multica-ai/andrej-karpathy-skills/create-feedback-loop-guideline-improvement) Learn to build an effective feedback loop for guideline improvement in LLM-assisted development. Capture failures, map principles, and validate fixes for better AI guidance. - Tags: how-to-guide - Published: 2026-04-19 ### [Debugging Techniques for Incorrect Guideline Application in Karpathy Coding Guidelines](/multica-ai/andrej-karpathy-skills/debugging-incorrect-guideline-application) Fix Karpathy coding guideline errors. Learn to debug plugin config, YAML front matter, and workflow integration issues for smoother development. Improve your application now. - Tags: debugging-guide - Published: 2026-04-19 ### [How to Adapt Guidelines to Different Programming Languages: Language-Agnostic Principles for LLM-Assisted Development](/multica-ai/andrej-karpathy-skills/guidelines-adapt-programming-languages) Learn how the Karpathy guidelines adapt to any programming language. Discover language-agnostic principles for effective LLM-assisted development, focusing on behavior not syntax. - Tags: best-practices - Published: 2026-04-19 ### [Real-World Performance Impact of Test-First Verification](/multica-ai/andrej-karpathy-skills/performance-impact-test-first-verification) Discover the real-world performance impact of test-first verification. Reduce latency spikes and CPU usage with this agile approach for measurable improvements. - Tags: testing - Published: 2026-04-19 ### [Handling "Simplicity First" Conflicts with Existing Complex Architecture](/multica-ai/andrej-karpathy-skills/simplicity-first-conflicts-architecture) Resolve 'Simplicity First' conflicts with complex architecture. Evaluate abstractions, apply surgical changes, and verify impact through goal-driven execution for cleaner systems. - Tags: best-practices - Published: 2026-04-19 ### [4 Common LLM Coding Mistakes the Karpathy Guidelines Prevent](/multica-ai/andrej-karpathy-skills/llm-coding-mistakes-guidelines-prevent) Avoid common LLM coding mistakes like fabricating APIs or over-engineering. The Karpathy Guidelines offer proven methods to build robust AI systems. - Tags: best-practices - Published: 2026-04-19 ### [How the Plugin System Handles Guideline Updates in the Andrej Karpathy Skills Repository](/multica-ai/andrej-karpathy-skills/plugin-system-handle-guideline-updates) Discover how the plugin system automatically handles guideline updates in the Andrej Karpathy Skills Repository. Simply edit SKILL.md to propagate changes to users. - Tags: how-to-guide - Published: 2026-04-19 ### [How to Override or Customize Guidelines for Specific Files in Cursor](/multica-ai/andrej-karpathy-skills/override-customize-guidelines-specific-files) Customize file guidelines in multica-ai/andrej-karpathy-skills. Disable global settings and use glob patterns in .cursor/rules to apply rules to specific files. - Tags: how-to-guide - Published: 2026-04-19 ### [Relationship Between Claude Code Skills and Cursor Rules: A Complete Guide](/multica-ai/andrej-karpathy-skills/claude-code-skills-cursor-rules-relationship) Explore the relationship between Claude Code skills and Cursor rules. Understand how these mirrored configurations enforce identical behavioral guidelines for developers. - Tags: deep-dive - Published: 2026-04-19 ### [How to Measure Guideline Impact on Code Quality: A Practical Guide to Karpathy-Inspired Metrics](/multica-ai/andrej-karpathy-skills/measure-guideline-impact-code-quality) Measure guideline impact on code quality using Karpathy inspired metrics. Track diff size, review requests, lint warnings, test failures and PR cycle time to see real improvements. - Tags: tutorial - Published: 2026-04-19 ### [What Happens When Test‑First Verification Finds Issues in Andrej Karpathy Skills](/multica-ai/andrej-karpathy-skills/test-first-verification-unexpected-issues) Discover what happens when test-first verification finds issues. Learn how developers fix bugs, iterate on code, and prevent regressions in the Andrej Karpathy skills repo. - Tags: deep-dive - Published: 2026-04-19 ### [How Surgical Changes Interact with Linting and Formatting Tools](/multica-ai/andrej-karpathy-skills/surgical-changes-interact-linting-formatting) Learn how Surgical Changes works with linting and formatting tools by scoping them to modified files. Keep intentional edits clear and prevent bulk reformatting. - Tags: how-to-guide - Published: 2026-04-19 ### [How to Create a Custom Skill for Claude Code Using the Karpathy Guidelines](/multica-ai/andrej-karpathy-skills/create-custom-skill-based-guidelines) Learn to create custom skills for Claude Code following Karpathy guidelines. Add a SKILL.md file, define metadata, write guidelines, and register your skill. - Tags: how-to-guide - Published: 2026-04-19 ### [Applying Karpathy-Inspired Guidelines in Monorepos with Multiple Projects](/multica-ai/andrej-karpathy-skills/apply-guidelines-monorepos-multiple-projects) Learn to apply Karpathy-inspired guidelines effectively within monorepos housing multiple projects. Streamline development workflows and ensure consistency across your codebase. - Tags: best-practices - Published: 2026-04-19 ### [How Goal-Driven Execution Reduces User Clarification Needs in LLM Workflows](/multica-ai/andrej-karpathy-skills/goal-driven-execution-reduce-clarification) Discover how Goal-Driven Execution transforms vague LLM instructions into clear success criteria, reducing user clarification needs and enabling autonomous validation. - Tags: deep-dive - Published: 2026-04-19 ### [Tradeoffs of Simplicity First in Large Codebases: A Practical Guide](/multica-ai/andrej-karpathy-skills/simplicity-first-tradeoffs-large-codebases) Explore the tradeoffs of simplicity first in large codebases. Learn how to balance minimal code with shared standards to avoid fragmentation and duplication in your projects. - Tags: best-practices - Published: 2026-04-19 ### [How Think Before Coding Handles Ambiguous Requirements in LLM Workflows](/multica-ai/andrej-karpathy-skills/think-before-coding-ambiguous-requirements) Learn how Think Before Coding addresses ambiguous requirements in LLM workflows by surfacing assumptions, enumerating interpretations, and seeking user clarification before coding. - Tags: deep-dive - Published: 2026-04-19 ### [How to Merge Project-Specific Guidelines with Base Karpathy Guidelines](/multica-ai/andrej-karpathy-skills/merge-project-guidelines-karpathy-guidelines) Merge Karpathy guidelines with project-specific rules by appending the core block to your CLAUDE.md and adding a custom section for unique guidelines. - Tags: best-practices - Published: 2026-04-19 ### [How to Debug When Claude Code Doesn't Follow Behavioral Guidelines](/multica-ai/andrej-karpathy-skills/debug-claude-code-behavioral-guidelines) Debug Claude Code ignoring guidelines by checking configuration, plugin installation, CLAUDE.md file presence, and cache freshness for quick fixes, not code bugs. - Tags: how-to-guide - Published: 2026-04-19 ### [How the Claude Code Plugin System Loads and Applies Skills from the Marketplace](/multica-ai/andrej-karpathy-skills/how-claude-code-plugin-loads-skills-marketplace) Explore how the Claude Code plugin system loads and applies skills. Discover the processes behind marketplace integration, plugin installation, and skill application for enhanced coding. - Tags: internals - Published: 2026-04-19