Agentapp Builder — Development agent for Claude Code
Agent 应用构建工程师(Context Engineering 合一).
How to install Agentapp Builder
Installs to ~/.claude/agents/xuanbingbingo-claude-agent-dev-team-agentapp-builder.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/xuanbingbingo/claude-agent-dev-team/HEAD/agents/agentapp-builder.md -o ~/.claude/agents/xuanbingbingo-claude-agent-dev-team-agentapp-builder.md Restart Claude Code, or start a new session, for it to be picked up.
What Agentapp Builder does
name: agentapp-builder description: Agent 应用构建工程师(Context Engineering 合一)。当需要编写 system prompt / 角色设定、定义工具(tool/function calling)契约、设计 RAG 检索与 context 拼装、或在 eval 循环里根据评估反馈迭代改进这些时激活。由 agent-dev-team 剧本在 Phase 3-5 调用。核心理念:prompt、工具、RAG、记忆本质是同一件事——往模型上下文里拼什么。严格按能力画像与架构 spec 实现,改动后必须交回 eval-lead 跑评估验证,禁止自评自夸。 tools: Read, Write, Edit, Bash, Glob, Grep model: opus
角色定义
你是 Agent 应用的构建工程师,专精 **Context Engineering**——决定"每一步往模型的上下文窗口里到底拼进什么"。你的核心认知:**system prompt、few-shot、工具定义、RAG 检索结果、记忆——这些
Alternatives in Development
- Agentlas Core Engine Meta Agent — Use this agent when the user asks for /meta-agent, a single agent builder, multi-agent team builder, or packag 1.1k ★
- Flask Route Builder — Use this agent to add, modify, or debug Flask routes in app.py 204 ★
- Chatforyou Lead — chatforyou-dev-team의 팀 리더 109 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Agentapp Eval Lead
Agent 应用 Eval 负责人(评估裁判,团队灵魂角色)。当需要定义"这个 agent 怎么算做对了""该度量哪几层""红线画在哪"、构建可自动跑的评估集、或在迭代中回归评估判断"这次改动是变好还是变坏"时激活。由
Agentapp Reality Checker
Agent 应用最终验收官。在 eval 循环收敛、安全审查完成后,由 agent-dev-team 剧本在 Phase 7 激活。对照能力画像与 eval 达标情况做整体验收,默认判决是"需要返工",只有压倒性证据才判
Context Builder
Builds architectural understanding before any verdict — runs Phase 0 setup and Phase 0.5 Context Reconstructio
Agent Tool Integration Engineer
Use this agent to wire tools and function-calling into an agent loop reliably — clean tool schemas, errors fed
Plugin Dev
Scoped guide for work on the herdr-plugin-pipeline plugin itself — the hpipe CLI, the supervisor, the phase ma
Red Team Reviewer
Fresh-eyes adversarial review of the skill itself (not its outputs). Use before calling an iteration done, and
Related Skills
Tools Comparison Table
A 107-topic deep dive into ChatGPT, Claude & Gemini—covering AI architecture, multimodality, prompting, contex
Secure AI Customer Support Platform
Secure LLM-powered customer support platform: an AI agent (Claude) with grounded RAG answers, authorization-en
AI Coding Exercises
Hands-on practice for AI engineering interviews: build LLM evals from scratch with the Claude API, plus drills