模型接力

model-handoff

by bwiley1989

Maintain a HANDOFF.md file in the workspace so context survives seamlessly when switching between LLM models (e.g. Claude → GPT → Gemini). Use when the user says they are switching models, asks how to preserve context across model switches, wants to save tokens by rotating models, or asks how a new model can pick up where the last one left off. Also use proactively during long sessions to keep HANDOFF.md current.

4.5kAI 与智能体未扫描2026年4月20日

安装

claude skill add --url https://github.com/openclaw/skills

文档

Model Handoff Skill

Enables seamless context continuity when switching between LLM models mid-session. Maintains a HANDOFF.md file that any model can read to instantly understand the current project state, active tasks, and behavioral expectations.

Core Concept

Every model starts a session cold. HANDOFF.md is a dense, always-current fast-boot file that eliminates the ramp-up. It is the single source of truth for model-to-model context transfer.

HANDOFF.md Structure

Write HANDOFF.md to the workspace root with these sections:

markdown
# HANDOFF.md — Model Switch Context

## Who you are
[Agent name, persona, tone, key behavioral rules. Reference SOUL.md if present.]

## Who you're helping
[User name, role, location, preferences. Reference USER.md and MEMORY.md if present.]

## Active projects
[For each project: name, status, key files, next steps. Be specific — include file paths.]

## Agent roster
[If multi-agent: list agent IDs, models, roles.]

## Key credentials & tools
[Point to credential files — never inline secrets. e.g. "Azure SP creds: azure-config.json"]

## Behavioral rules
[Critical rules a new model must follow. Keep to essentials only.]

## How to keep this file current
[Brief note on when to update.]

## Last updated
[Timestamp + 1-line session summary]

When to Create/Update

Create HANDOFF.md when:

  • User asks about switching models
  • User asks how to preserve context
  • HANDOFF.md does not yet exist

Update HANDOFF.md when:

  • User says "switching to [model]" — update immediately before they go
  • User says "update HANDOFF" or "log everything"
  • A significant project milestone is reached (new project started, major decision made, new agent added)
  • The session has been running for several hours with significant new context

Keep current proactively — do not wait to be asked. Update during long sessions when meaningful things happen.

Wiring HANDOFF.md into the Workspace

After creating HANDOFF.md, add a reference in AGENTS.md so every model is instructed to read it on a model switch:

markdown
## Every Session
...
- **If you are a new model taking over** (model switch): Read `HANDOFF.md` first — it's your fast-boot summary of everything active

What to Say When Switching

Tell the user to open with this when switching to a new model:

"Read HANDOFF.md. You are [agent name]."

That single line forces the new model to self-load before responding.

Writing Guidelines

  • Dense, not verbose — every line earns its place
  • File paths, not descriptions — "see azure-config.json" not "credentials are stored somewhere"
  • Never inline secrets — point to credential files only
  • No personal/private data — HANDOFF.md may be shared; keep sensitive context in MEMORY.md
  • Remove stale content — delete completed projects and outdated context on each update
  • Last updated timestamp — always include so the receiving model knows how fresh it is

References

  • See references/template.md for a copy-paste HANDOFF.md starter template

相关 Skills

Claude接口

by anthropics

Universal
热门

面向接入 Claude API、Anthropic SDK 或 Agent SDK 的开发场景,自动识别项目语言并给出对应示例与默认配置,快速搭建 LLM 应用。

想把Claude能力接进应用或智能体,用claude-api上手快、兼容Anthropic与Agent SDK,集成路径清晰又省心

AI 与智能体
未扫描134.4k

RAG架构师

by alirezarezvani

Universal
热门

聚焦生产级RAG系统设计与优化,覆盖文档切块、检索链路、索引构建、召回评估等关键环节,适合搭建可扩展、高准确率的知识库问答与检索增强应用。

面向RAG落地,把知识库、向量检索和生成链路系统串联起来,做架构设计时更清晰,也更少踩坑。

AI 与智能体
未扫描14.9k

提示工程专家

by alirezarezvani

Universal
热门

覆盖Prompt优化、Few-shot设计、结构化输出、RAG评测与Agent工作流编排,适合分析token成本、评估LLM输出质量,并搭建可落地的AI智能体系统。

把提示优化、LLM评测到RAG与智能体设计串成一套方法,适合想系统提升AI开发效率的人。

AI 与智能体
未扫描14.9k

相关 MCP 服务

顺序思维

编辑精选

by Anthropic

热门

Sequential Thinking 是让 AI 通过动态思维链解决复杂问题的参考服务器。

这个服务器展示了如何让 Claude 像人类一样逐步推理,适合开发者学习 MCP 的思维链实现。但注意它只是个参考示例,别指望直接用在生产环境里。

AI 与智能体
85.7k

知识图谱记忆

编辑精选

by Anthropic

热门

Memory 是一个基于本地知识图谱的持久化记忆系统,让 AI 记住长期上下文。

帮 AI 和智能体补上“记不住”的短板,用本地知识图谱沉淀长期上下文,连续对话更聪明,数据也更可控。

AI 与智能体
85.7k

PraisonAI

编辑精选

by mervinpraison

热门

PraisonAI 是一个支持自反思和多 LLM 的低代码 AI 智能体框架。

如果你需要快速搭建一个能 24/7 运行的 AI 智能体团队来处理复杂任务(比如自动研究或代码生成),PraisonAI 的低代码设计和多平台集成(如 Telegram)让它上手极快。但作为非官方项目,它的生态成熟度可能不如 LangChain 等主流框架,适合愿意尝鲜的开发者。

AI 与智能体
7.7k

评论