Agent Builder

Build, review, troubleshoot, optimize, and extend AI agents — single-agent, multi-agent, RAG, and production systems. Covers framework selection, pattern selection (topology, behavioral, data flow), agent orchestration, state design, and tool design. Defaults to Python with LangChain/LangGraph; also supports CrewAI, Strands, OpenAI Agents SDK, Google ADK, and Mastra. Use whenever the user wants to build, design, scaffold, review, debug, optimize, extend, or migrate an agent or agentic system.

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add NikolasMarkou/agent-builder --skill agent-builder
Manual install
git clone --depth 1 https://github.com/NikolasMarkou/agent-builder.git /tmp/agent-builder
mkdir -p ~/.claude/skills
cp -r /tmp/agent-builder/src ~/.claude/skills/agent-builder

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Agent Builder

Build production-grade AI agents from requirements. Default stack: Python + LangChain v1.2.x + LangGraph v1.0.x.

## Design Axioms

Six principles that recur across every reference file. Apply them at every workflow step -- they are the difference between a demo and a production system. Most impactful for Moderate complexity and above.

| Axiom | Rule | Violated When |
|---|---|---|
| **Tiered escalation** | Cheap/deterministic first, LLM only for judgment, human as backstop | You send every query through the LLM when 60% could be a DB lookup |
| **Decompose** | Split complex tasks into simple, independent pieces | You ask an LLM to score 5 criteria in one call instead of 5 binary questions |
| **Model costs first** | Calculate token math at expected scale before writing code | You discover at launch that your $500/mo PoC costs $847K/mo in production |
| **Minimize context** | Send the minimum data needed; format matters (15-20pp accuracy swing) | You dump 2000-line files into context when 20 lines were relevant |
| **Calibrate on real data** | Build evals from real failures, not synthetic examples | Your benchmarks pass but production users hit edge cases you never tested |
| **Document failure modes** | Every pattern has known failures -- catalog and mitigate them upfront | Your loop runs forever because you didn't set a max iteration cap |
Read the full skill on GitHub

Details

Skill name
agent-builder
Path
src/SKILL.md
License
GPL-3.0
Repo updated
8 June 2026

FAQ

What does the Agent Builder skill do?

Build, review, troubleshoot, optimize, and extend AI agents — single-agent, multi-agent, RAG, and production systems. Covers framework selection, pattern selection (topology, behavioral, data flow), agent orchestration, state design, and tool design. Defaults to Python with LangChain/LangGraph; also supports CrewAI, Strands, OpenAI Agents SDK, Google ADK, and Mastra. Use whenever the user wants to build, design, scaffold, review, debug, optimize, extend, or migrate an agent or agentic system.

How do I install the Agent Builder skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add NikolasMarkou/agent-builder --skill agent-builder. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Agent Builder popular?

It lives in NikolasMarkou/agent-builder, which has 5 GitHub stars and 0 forks.

Related skills

Agent Design Fundamentals

10

Archive228/lab-skills

Decision rules for whether to build an LLM agent and how to architect it, distilled from OpenAI's "A Practical Guide to Building Agents": build/no-build criteria, model selection, tool design, single-agent vs multi-agent orchestration (manager pattern vs decentralized handoffs), layered guardrails, and human-in-the-loop triggers. Use when designing, reviewing, or refactoring an agent system — before choosing an architecture, adding more agents, defining tools, or shipping an agent to production without guardrails.

AI, LLMs & Agents

Storyblok Use MCP

9

storyblok/agent-plugin

Use when interacting with Storyblok through the MCP server — the tool model, operation names and payload shapes for stories, assets, components, datasources and every other resource. Load it before calling `mcp__storyblok__` tools, alongside the skill that owns the task.

AI, LLMs & Agents

Agent Evals

10

Archive228/lab-skills

Build and maintain evaluation suites for AI agents: bootstrap 20-50 tasks from real failures, pick code-based vs model-based vs human graders by agent architecture (coding, conversational, research, computer-use), write unambiguous tasks with reference solutions, and keep suites alive via ownership and transcript reading. Use when creating evals from zero, adding graders or eval cases, debugging suspicious eval scores, or deciding pass@k vs pass^k metrics for an agent product.

AI, LLMs & Agents

Recover Semantic Model

9

kr4chinin/skills

Recover a program's minimal semantic model (the state that changes what users see, how its transitions repeat and reorder, what each effect requires). Use when asked to recover or audit that model, to check what is idempotent or order-dependent, or to split a refactor into provably equivalent and "trust me" changes.

AI, LLMs & Agents

Agent Failure Preflight

10

Archive228/lab-skills

Pre-deployment failure-mode gate for agentic AI systems, distilled from Microsoft AI Red Team's Taxonomy of Failure Modes in Agentic AI Systems v2.0 (grounded in 12 months of red-team engagements). Run it before shipping or materially changing an agent — adding tools/MCP servers/plugins, enabling persistent memory, adding sub-agents or orchestration, enabling computer use, or expanding permissions — to verify concrete controls against memory poisoning, HitL bypass, goal hijacking, supply-chain compromise, session context contamination, capability disclosure, and multi-agent cascade failures.

AI, LLMs & Agents

AI Paper Writing

9

HOWILLMAKEIT/skills

辅助 AI 论文写作,分两类任务。一是论文主图、方法图、机制图和图形摘要:搜集顶会同类型参考图,写网页生图提示词,用户手动生图后重建为原生可编辑 PPT。二是摘要、Introduction、实验等文字段落:先搜集相关领域顶会的对应段落到文件夹(md),暂停;用户需要时再阅读并总结写法与共同逻辑,暂停;再按总结的逻辑顺序,结合用户提供的代码和论文草稿完成写作。文字写作的三步各自暂停,等用户确认后才进入下一步。实测数据图应从真实数据绘制。

AI, LLMs & Agents

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.