206 Prompt Patterns for AI Agent Development
206 个 AI Agent 提示词模式
Extracted from 500+ production AI agent plugins, plus the superpowers, claude-plugins-official, and mattpocock/skills repositories. Browse by category, technique, or standard.
从 500+ 生产 AI agent 插件，以及 superpowers、claude-plugins-official、mattpocock/skills 仓库中提炼。按分类、技巧或标准浏览。
🔍 Interactive pattern browser → — search + filter all 206 by category / source / confidence, EN & 中文.
Patterns 156–206 are the 2026-07 harvest; full write-ups in progress.
Full Index / 总索引
Categories / 分类
| Category / 分类 |
EN |
中文 |
| Structural Scaffolding / 结构脚手架 |
Read |
阅读 |
| Input/Output Contracts / 输入输出契约 |
Read |
阅读 |
| Execution Control / 执行控制 |
Read |
阅读 |
| Knowledge & Context / 知识与上下文 |
Read |
阅读 |
| Agent Orchestration / Agent 编排 |
Read |
阅读 |
| Safety & Trust / 安全与信任 |
Read |
阅读 |
| Quality & Feedback / 质量与反馈 |
Read |
阅读 |
| Advanced I/O & Domain / 高级 I/O |
Read |
阅读 |
| Advanced Orchestration / 高级编排 |
Read |
阅读 |
| Advanced Quality / 高级质量 |
Read |
阅读 |
| Advanced Safety / 高级安全 |
Read |
阅读 |
| Advanced Workflow / 高级工作流 |
Read |
阅读 |
| Gap Fills / 补充模式 |
Read |
阅读 |
| Open Source Skills / 开源技能 |
Read |
阅读 |
| Karpathy Behavioral / Karpathy 行为 |
Read |
阅读 |
| Claude Code Platform |
Read |
阅读 |
| Claude Code Platform (Extended) |
Read |
阅读 |
Techniques / 技巧
| Topic / 主题 |
EN |
中文 |
| Token-Level Techniques (9) / Token 级技巧 |
Read |
阅读 |
| Anti-Laziness Strategies (8) / 防偷懒策略 |
Read |
阅读 |
| Skill Architecture / 技能架构 |
Read |
阅读 |
| Branching Stability / 分支稳定性 |
Read |
阅读 |
| Reference-Skip Playbook / 引用跳过应对 |
Read |
阅读 |
| Good vs Bad Template / 好坏对比模板 |
Read |
阅读 |
Standards / 标准
| Topic / 主题 |
EN |
中文 |
| Quality Standards (P0/P1/P2) / 质量标准 |
Read |
阅读 |
| Review Checklist (9 dimensions) / 评审清单 |
Read |
阅读 |