with one click
prompt-engineering
提示工程技术与 Agent 设计模式
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Menu
提示工程技术与 Agent 设计模式
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
研究一个**已知的** GitHub 项目/repo,判断它对 LiYe Systems 的进化有没有可借鉴之处,产出定型的 evolution-recon 报告并归档到 recon-log。Use whenever the user gives a specific GitHub repo URL and asks to 研究/研读/消化/digest/study it, wants to know whether a project 对 LiYe / AGE / loamwise 有没有帮助/可借鉴/值得参考, asks "这个项目对我们有没有用", floats "看看 <repo> 有没有能学的", or types /github-digest — even if they never name the report format. This is the KNOWN-repo research step that sits UPSTREAM of github-scout (which DISCOVERS unknown repos) and source-intake (which INGESTS a chosen repo). Reach for it when the repo is already chosen and the real question is "is it worth our attention / what can LiYe learn". Do NOT use it to search/discover repos (that is github-scout) or to actually vendor/ingest a repo (that is source-intake).
Use when testing SFC v0.2 frontmatter compatibility.
Claude Skills 创建指南与最佳实践
SFC patched skill. Use when "ui-ux" is relevant.
内容研究与写作助手
视觉设计与艺术创作工具
| name | prompt-engineering |
| description | 提示工程技术与 Agent 设计模式 |
| domain | 00_Core_Utilities |
| category | meta |
| version | 1.0.0 |
| status | active |
| source | awesome-claude-skills |
| source_url | https://github.com/ComposioHQ/awesome-claude-skills |
| license | Apache-2.0 |
| skeleton | reference |
| triggers | {"commands":["/prompt-engineering"],"patterns":["prompt-engineering"]} |
| inputs | {"required":[],"optional":[]} |
| outputs | {"artifacts":["SKILL.md"]} |
| failure_modes | [{"symptom":"Missing required inputs or context","recovery":"Provide the missing info and retry"},{"symptom":"Unexpected tool/runtime failure","recovery":"Rerun with minimal steps; escalate after 3 failures"}] |
| verification | {"evidence_required":true,"how_to_verify":["node .claude/scripts/sfc_lint.mjs <skill_dir>"]} |
| governance | {"constitution":"_meta/governance/SKILL_CONSTITUTION_v0.1.md","policy":"_meta/policies/DEFAULT_SKILL_POLICY.md"} |
来源: ComposioHQ/awesome-claude-skills 适配: LiYe OS 三层架构
教授提示工程技术,包括 Anthropic 最佳实践、Agent 设计模式、上下文工程。
当需要优化 AI 交互时:
# System Prompt 结构
## 角色定义
你是 [角色],专长于 [领域]。
## 核心指令
1. [首要任务]
2. [次要任务]
## 约束条件
- [限制 1]
- [限制 2]
## 输出格式
[期望的输出格式说明]
## 示例
[Few-shot 示例]
| 技术 | 描述 | 效果 |
|---|---|---|
| XML 标签 | 使用 <tag> 结构化内容 | 提升解析准确度 |
| Chain of Thought | 引导逐步推理 | 提升复杂任务表现 |
| 角色扮演 | 明确定义 AI 角色 | 输出更一致 |
| Few-shot | 提供示例 | 输出格式更稳定 |
| 显式约束 | 明确禁止行为 | 减少意外输出 |
┌─────────────────────────────────┐
│ System Prompt │ ← 角色、规则、格式
├─────────────────────────────────┤
│ Retrieved Context │ ← RAG 检索内容
├─────────────────────────────────┤
│ Conversation History │ ← 对话历史
├─────────────────────────────────┤
│ User Message │ ← 用户输入
└─────────────────────────────────┘
ReAct 模式:
Thought: 我需要...
Action: [工具调用]
Observation: [结果]
Thought: 根据结果...
Action: [下一步]
...
Final Answer: [最终答案]
Plan-and-Execute:
1. 分析任务 → 生成计划
2. 逐步执行计划
3. 根据反馈调整
4. 汇总结果
Multi-Agent 协作:
Coordinator → Researcher → Writer → Reviewer
↑__________________________________|
用户: 帮我设计一个代码审查 Agent 的 System Prompt
Claude: [使用 prompt-engineering 设计角色、规则、输出格式]
用户: 这个提示效果不好,帮我优化
Claude: [使用 prompt-engineering 分析问题、应用技术、重写提示]
用户: 我想让 Claude 自动完成研究任务
Claude: [使用 prompt-engineering 设计 ReAct 循环、定义工具、编排流程]
无外部依赖,纯方法论技能。
此技能被以下业务域引用:
本技能直接服务于 LiYe OS 的核心交互层:
.claude/packs/ 的 Context PacksCreated: 2025-12-28 | Adapted for LiYe OS