SkillAgentSearch skills...

The 15 Best Claude Code Skills for Building AI Apps

Ranked by quality, maintenance and a safety scan of every file · Updated

Skills for LLM features: prompts, RAG, embeddings and agent SDKs. We compared 17 Claude Code Skills for Building AI Apps that pass our quality bar, 94% of which came up clean on our safety scan; these are the top 15, no more than three from any one repository.

  1. Add a persistent wiki knowledge base to a NanoClaw group. Based on Karpathy's LLM Wiki pattern. Triggers on "add wiki", "wiki", "knowledge base", "llm wiki", "karpathy wiki".

    Quality 93/100Safety scan clean★ 31,114Updated 2 days ago

    Why it is here: updated 2 days ago; clean safety scan; 31,114 GitHub stars.

    npx skills add nanocoai/nanoclaw --skill add-karpathy-llm-wiki
  2. Overview of six AI engineering workflows: prompt evaluation (8-dimension scoring), context budget planning, RAG pipeline design, agent security audit (65-point checklist), eval harness building, and product sense coaching.

    Quality 97/100Safety scan clean★ 49,671Updated 3 days ago

    Why it is here: highest quality score in this list; updated 3 days ago; clean safety scan.

    npx skills add sickn33/agentic-awesome-skills --skill ai-engineering-toolkit
  3. 3

    agentica-prompts

    by parcadei

    Write reliable prompts for Agentica/REPL agents that avoid LLM instruction ambiguity

    Quality 86/100Safety scan clean★ 3,943Updated 2026-01-26

    Why it is here: clean safety scan; 3,943 GitHub stars.

    npx skills add parcadei/Continuous-Claude-v3 --skill agentica-prompts
  4. 4

    llm-gate

    by rohitg00

    LLM-powered quality verification using prompt hooks. Validates commit messages, code patterns, and conventions using AI before allowing operations. Use to set up intelligent guardrails.

    Quality 90/100Safety scan clean★ 2,876Updated 53 days ago

    Why it is here: clean safety scan; 2,876 GitHub stars.

    npx skills add rohitg00/pro-workflow --skill llm-gate
  5. 5

    Finetune a pretrained KERMT encoder on a labeled CSV. Validate the checkpoint and data, prepare features, and run containerized training. Use a local checkpoint or optionally download a pinned Hugging Face model bundle using HF_TOKEN if configured.

    Quality 92/100Safety scan clean★ 3,421Updated 54 days ago

    Why it is here: clean safety scan; 3,421 GitHub stars.

    npx skills add NVIDIA/skills --skill kermt-finetune
  6. 6

    llm-wiki-setup

    by daymade

    Co-creates a personal investment-research LLM Wiki (Karpathy's pattern) by interviewing the user for THEIR OWN analysis framework, not a template — a living CLAUDE.md of pure markdown + wikilinks, no RAG in place of compilation.

    Quality 87/100Safety scan clean★ 1,426Updated 53 days ago

    Why it is here: clean safety scan; 1,426 GitHub stars.

    npx skills add daymade/claude-code-skills --skill llm-wiki-setup
  7. AI skill for OpenClaw & Claude Code — recommend from 10000+ Nano Banana Pro (Gemini) image prompts. Smart search by use case, content remix, sample images.

    Quality 92/100Safety scan clean★ 1,912Updated today

    Why it is here: updated in the last day; clean safety scan; 1,912 GitHub stars.

    npx skills add YouMind-OpenLab/nano-banana-pro-prompts-recommend-skill
  8. AI 学习记录与成长追踪工具。用于记录 AI/LLM 学习笔记、使用心得、Prompt 技巧、工具体验等,并提供学习指导和规划。当用户提到以下任何话题时都应使用此 skill:AI 学习记录、学习笔记、AI 使用心得、Prompt 工程学习、模型对比体验、AI 工具使用记录、LLM 学习、RAG 学习、Agent 学习、MCP 学习、AI 微调实践、AI 学习规划、怎么学 AI、AI 入门、学习路线推荐、从零学 AI、推荐学习资源、AI 学习方向、如何系统学习 AI、学习方法建议、回顾学习记录、学习总结、AI 知识整理。即使用户只是随口提了一句\"今天试了一下 Claude\"或\"学了个新的…

    Quality 88/100Safety scan clean★ 2,149Updated 2026-07-20

    Why it is here: clean safety scan; 2,149 GitHub stars.

    npx skills add LeoYeAI/openclaw-master-skills --skill ai-learning-journal
  9. 9

    prompt-optimizer

    by getsentry

    Creates, optimizes, and iteratively refines agent prompts, system prompts, developer prompts, and reusable prompt templates

    Quality 81/100Safety scan clean★ 1,004Updated 56 days ago

    Why it is here: clean safety scan; 1,004 GitHub stars.

    npx skills add getsentry/skills --skill prompt-optimizer
  10. 10

    llm-icon-finder

    by daymade

    Finding and accessing AI/LLM model brand icons from lobe-icons library

    Quality 86/100Safety scan clean★ 1,426Updated 53 days ago

    Why it is here: clean safety scan; 1,426 GitHub stars.

    npx skills add daymade/claude-code-skills --skill llm-icon-finder
  11. A skill to claude code that enables brainstorming with other LLMs (ChatGPT, Gemini) before presenting the implementation plan to the user

    Quality 81/100Safety scan clean★ 506Updated 2026-01-08

    Why it is here: clean safety scan.

    npx skills add gcpdev/llm-council-skill
  12. 12

    aws-sdk-java-v2-bedrock

    by giuseppe-trisciuoglio

    Provides Amazon Bedrock patterns using AWS SDK for Java 2.x. Invokes foundation models (Claude, Llama, Titan), generates text and images, creates embeddings for RAG, streams real-time responses, and configures Spring Boot integration

    Quality 91/100Safety scan clean★ 370Updated 2026-09-10

    Why it is here: clean safety scan.

    npx skills add giuseppe-trisciuoglio/developer-kit --skill aws-sdk-java-v2-bedrock
  13. 13

    change-core-self

    by syahiidkamil

    Interview Boss about the project, then reason from first principles to design the ideal ATLAS operating identity/system-prompt for it — free to drop KISS/YAGNI/DRY/clean-architecture entirely when the project (and the LLM's own distribution) calls for a different mindset

    Quality 86/100Safety scan clean★ 406Updated 15 days ago

    Why it is here: updated 15 days ago; clean safety scan.

    npx skills add syahiidkamil/Software-Engineer-AI-Agent-Atlas
  14. 14

    llm-wiki-plugin

    by praneybehl

    Andrej Karpathy's LLM Wiki pattern as a skill & Claude Code plugin — turn accumulated sources into a self-maintaining, scalable markdown knowledge base.

    Quality 89/100Safety scan clean★ 124Updated 18 days ago

    Why it is here: updated 18 days ago; clean safety scan.

    npx skills add praneybehl/llm-wiki-plugin
  15. 15

    llm-data-automation

    by datadrivenconstruction

    Automate construction data processing using LLM (ChatGPT, Claude, LLaMA). Generate Python/Pandas scripts, extract data from documents, and create automated pipelines without deep programming knowledge.

    Quality 89/100Safety notes★ 333Updated 2026-08-22

    Why it is here: safety notes worth reading before install.

    npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill llm-data-automation

Side by side

#NameQualitySafetyStarsUpdated
1add-karpathy-llm-wiki93Clean31,1142 days ago
2ai-engineering-toolkit97Clean49,6713 days ago
3agentica-prompts86Clean3,9432026-01-26
4llm-gate90Clean2,87653 days ago
5kermt-finetune92Clean3,42154 days ago
6llm-wiki-setup87Clean1,42653 days ago
7nano-banana-pro-prompts-recommend-skill92Clean1,912today
8ai-learning-journal88Clean2,1492026-07-20
9prompt-optimizer81Clean1,00456 days ago
10llm-icon-finder86Clean1,42653 days ago
11llm-council-skill81Clean5062026-01-08
12aws-sdk-java-v2-bedrock91Clean3702026-09-10
13change-core-self86Clean40615 days ago
14llm-wiki-plugin89Clean12418 days ago
15llm-data-automation89Notes3332026-08-22

How we rank

Only skills we have read make the list: the file itself was fetched, it has a real description, it scores at least 55 of 100 on quality, its source still exists, its repository has at least 20 GitHub stars, and our safety scan found nothing high-risk. The order then follows adoption (GitHub stars, on a log scale), how squarely it fits this topic, and its quality score (how substantial and well structured the file is and how recently it changed), moved up for a clean safety scan and for real installs through the SkillAgent MCP server, and down for safety notes. At most three come from any one repository, and tools sold on evading bot detection are never listed. We do not rank by install counts reported by a command-line tool: they are easy to inflate, and our research found repositories with no stars claiming hundreds of thousands of installs. The list is recomputed on every visit.

Questions

What is the best Claude Code Skill for Building AI Apps?

add-karpathy-llm-wiki by nanocoai ranks first: updated 2 days ago, clean safety scan, 31,114 GitHub stars. The full ranking above lists 15 alternatives with their trade-offs.

How do I install one?

Each entry shows a one-line command. `npx skills add` installs a skill for Claude Code, Cursor, Codex, Gemini CLI and other agents that read skill files.

Are these safe to install?

Every file is scanned for instructions that work against you: hidden text, data sent to outside servers, disabled approval prompts, leaked keys. Skills with a warning are never listed here, and 94% of the 17 we compared for this guide came up clean. Read a skill before installing it all the same: it runs with your permissions.