Skip to content
CodexWork
Any AIChatGPTClaudeCodexCopilotGeminiGrokOther agents

Use materials in your tool

All tool notes →

Hand a project to a fresh agent

Give it the project's state. Get a handoff file a new agent can pick up.

0608
Designed method · documented tool support

No recorded run of this exact instruction yet. Official sources document relevant tool support; the task method and synthetic sample are authored examples, not vendor-tested procedures.

Separate chats do not automatically share context, files or permissions.

Read the full instruction ↓

What you supply

  • Authorized project files and current work record
  • Scope, desired outcome and the receiver’s known access

What it saves

  • Self-contained handoff file
  • Source and validation references
  • Next action with stop conditions

Designed example · synthetic data

See the finished example

Read the complete example →Starting materials →Editable Markdown ↓

The instruction

Prepare a handoff from the authorized files I provide so a fresh agent can continue without this chat. Use the project’s current files as evidence; distinguish confirmed facts, judgments and unknowns. Do not assume the receiver can see local paths, credentials, browser sessions or attachments unless its access is confirmed.

Save handoff.md with the objective, current state, authoritative source paths/links, completed changes, actual validation results, unresolved decisions, next action, expected output and stop conditions. Include only history that changes the receiver’s work. Mark unrun checks as unrun and proposals as proposals. Describe how to obtain needed inputs without granting access or exposing secrets.

Check the handoff against its source files. If an independent fresh receiver is available and its use is authorized, let it attempt the next action using only the handoff and accessible inputs, and record what it actually did. Otherwise label the check “Source check only; fresh receiver not tested.” Do not simulate a receiver and call that independent proof. Do not send the handoff to another chat/person, run the next task or grant access unless separately authorized. Return the saved handoff and remaining access gaps.

Show plain text
Prepare a handoff from the authorized files I provide so a fresh agent can continue without this chat. Use the project’s current files as evidence; distinguish confirmed facts, judgments and unknowns. Do not assume the receiver can see local paths, credentials, browser sessions or attachments unless its access is confirmed.

Save handoff.md with the objective, current state, authoritative source paths/links, completed changes, actual validation results, unresolved decisions, next action, expected output and stop conditions. Include only history that changes the receiver’s work. Mark unrun checks as unrun and proposals as proposals. Describe how to obtain needed inputs without granting access or exposing secrets.

Check the handoff against its source files. If an independent fresh receiver is available and its use is authorized, let it attempt the next action using only the handoff and accessible inputs, and record what it actually did. Otherwise label the check “Source check only; fresh receiver not tested.” Do not simulate a receiver and call that independent proof. Do not send the handoff to another chat/person, run the next task or grant access unless separately authorized. Return the saved handoff and remaining access gaps.

Stops in this kit

06

Sending the handoff or running the next task

08

Unresolved decisions for the receiver

Access the receiver needs (the file grants none)

Checks to run

  1. The next action is possible from the supplied handoff and accessible files.
  2. Actual validation, unrun checks and open decisions are distinct.
  3. Private data and inaccessible machine paths are not mistaken for usable shared inputs.
  4. A source check is not labeled an independent receiver test.
Open the checklist →

What has been tested

Designed method · documented tool support

No recorded run of this exact instruction yet. Official sources document relevant tool support; the task method and synthetic sample are authored examples, not vendor-tested procedures.

Checked 2026-10-13

Limits

  • Separate chats do not automatically share context, files or permissions.
  • A handoff cannot transfer credentials or computer access by describing them.
  • Creating a new chat, messaging it and executing its next task require actual authorization.

Optional tool setup

Using a kit

  1. Provide the current work record and the files the next agent can actually access.
  2. Generate a handoff focused on the next concrete action.
  3. Check sources; use a separately authorized fresh receiver when available.
Use a kit in your tool →

ChatGPT

  1. Paste the instruction, attach the sample files or your own, and read the stops first.
  2. Pick Work for a multi-step task and supply the context it needs.
  3. For reuse, follow the documented skill or plugin setup on your surface. Downloading a kit does not install it.

Check the permissions for the tools and actions you use. The kit's stops are instructions to the agent; they do not create an enforced approval setting.

ChatGPT · dated notes →

Claude

  1. Paste the instruction, attach the sample files or your own, and read the stops first.
  2. In Cowork, choose the authorized folder containing your inputs where that feature is available. A generic kit-to-skill import was not established by the reviewed pages.
  3. In Claude Code, use CLAUDE.md for project guidance. AGENTS.md is read by default on supported versions (2.1.277+) only when project-path CLAUDE.md or CLAUDE.local.md does not take precedence; session and configuration exceptions apply.

Check the controls in your Claude surface. Anthropic describes approval handbacks, but the kit's stops do not enforce a product permission boundary.

Claude · dated notes →

Codex

  1. Paste the instruction, attach the sample files or your own, and read the stops first.
  2. For project guidance, use AGENTS.md. For a reusable skill, follow the supported setup, such as .agents/skills/<name>/SKILL.md in a repository.
  3. Check the sandbox and approval settings before you start the job.

Codex has sandbox and approval settings. Check them before a job that touches real files.

Codex · dated notes →

Copilot

  1. Paste the plain instruction into a Copilot chat you can access, and read the stops first.
  2. Check which file inputs and actions your tenant's actual surface supports before supplying files.
  3. Keep the kit as a manual prompt. The reviewed posts do not document importing it as a skill, plugin or agent.

Ask your administrator which actions, permissions and approval controls your organization has enabled. A kit's stops do not create those controls.

Copilot · dated notes →

Gemini

  1. Paste the instruction, attach the sample files or your own, and read the stops first.
  2. Where Skills is available, copy or import the skill using that surface's documented setup.
  3. Gemini app and Workspace skills do not sync. Set up the skill separately on each available surface you use.

Some Gemini Tasks may require confirmation. Check current eligibility and permissions before real account work; the kit's stops are instructions, not enforced approval settings.

Gemini · dated notes →

Grok

  1. Paste the plain instruction as a message and read the stops first.
  2. If Bot is available to you, show it the workflow; xAI describes routine learning by following along.
  3. Keep the instruction as plain text; we could not confirm SKILL.md support.

xAI describes approval handbacks for Grok Bot. Check what your account and connected apps allow before real actions.

Grok · dated notes →

Other agents

  1. Paste the plain instruction into an agent that accepts text, then supply inputs through its supported route and read the stops first.
  2. If your agent reads AGENTS.md, save the instruction there.
  3. If the named client supports Agent Skills, use the documented skill folder location and setup for that client and version.

This varies by agent. Look up its approval settings; the kit's stops stay in the instruction wherever you paste it.

Other agents · dated notes →
Markdown · README.md
# Hand a project to a fresh agent

Give it the project's state. Get a handoff file a new agent can pick up.

## Start

1. Edit [brief.md](agent-handoff/files/brief.md) and [input-template.md](agent-handoff/files/input-template.md).
2. Supply authorized inputs to your own file-capable agent with [instruction.txt](agent-handoff/files/instruction.txt).
3. Inspect the saved outputs using [checks.md](agent-handoff/files/checks.md) before using them. Keep original inputs.

## Inputs

- Authorized project files and current work record
- Scope, desired outcome and the receiver’s known access

## Outputs

- Self-contained handoff file
- Source and validation references
- Next action with stop conditions

## Steps

- Provide the current work record and the files the next agent can actually access.
- Generate a handoff focused on the next concrete action.
- Check sources; use a separately authorized fresh receiver when available.

## Instruction

Prepare a handoff from the authorized files I provide so a fresh agent can continue without this chat. Use the project’s current files as evidence; distinguish confirmed facts, judgments and unknowns. Do not assume the receiver can see local paths, credentials, browser sessions or attachments unless its access is confirmed.

Save handoff.md with the objective, current state, authoritative source paths/links, completed changes, actual validation results, unresolved decisions, next action, expected output and stop conditions. Include only history that changes the receiver’s work. Mark unrun checks as unrun and proposals as proposals. Describe how to obtain needed inputs without granting access or exposing secrets.

Check the handoff against its source files. If an independent fresh receiver is available and its use is authorized, let it attempt the next action using only the handoff and accessible inputs, and record what it actually did. Otherwise label the check “Source check only; fresh receiver not tested.” Do not simulate a receiver and call that independent proof. Do not send the handoff to another chat/person, run the next task or grant access unless separately authorized. Return the saved handoff and remaining access gaps.

## Decisions that need your context

- Unresolved decisions for the receiver
- Access the receiver needs (the file grants none)
- Sending the handoff or running the next task

## Settled by the brief

No separately annotated settled rules are supplied. Follow the instruction as written.

## Check the output

- The next action is possible from the supplied handoff and accessible files.
- Actual validation, unrun checks and open decisions are distinct.
- Private data and inaccessible machine paths are not mistaken for usable shared inputs.
- A source check is not labeled an independent receiver test.

## Limits

- Separate chats do not automatically share context, files or permissions.
- A handoff cannot transfer credentials or computer access by describing them.
- Creating a new chat, messaging it and executing its next task require actual authorization.

## Materials

- [examples/agent-handoff/designed-example.html](agent-handoff/files/examples/agent-handoff/designed-example.html)
- [examples/agent-handoff/designed-example.md](agent-handoff/files/examples/agent-handoff/designed-example.md)
- [examples/agent-handoff/input.md](agent-handoff/files/examples/agent-handoff/input.md)

## Evidence

[Recorded context and file identities](agent-handoff/files/evidence.json)

Designed method · documented tool support

No recorded run of this exact instruction yet. Official sources document relevant tool support; the task method and synthetic sample are authored examples, not vendor-tested procedures.

Observed or source-checked: 2026-10-13.

## Revision and corrections

Revision: 2.0.1. Lifecycle: active.

[Current kit](https://codexwork.com/kits/agent-handoff/) · [Changes and corrections](https://codexwork.com/changes/)

[Rights and notices](agent-handoff/files/LICENSES.md) · [Sources](agent-handoff/files/sources.md) · [Manifest](agent-handoff/files/manifest.json)

Downloading these files does not run an agent, install a skill, create a schedule, or grant file or account permissions.
Open Markdown file ↗
SKILL.md
---
name: agent-handoff
description: "Give it the project's state. Get a handoff file a new agent can pick up."
---

# Hand a project to a fresh agent

Prepare a handoff from the authorized files I provide so a fresh agent can continue without this chat. Use the project’s current files as evidence; distinguish confirmed facts, judgments and unknowns. Do not assume the receiver can see local paths, credentials, browser sessions or attachments unless its access is confirmed.

Save handoff.md with the objective, current state, authoritative source paths/links, completed changes, actual validation results, unresolved decisions, next action, expected output and stop conditions. Include only history that changes the receiver’s work. Mark unrun checks as unrun and proposals as proposals. Describe how to obtain needed inputs without granting access or exposing secrets.

Check the handoff against its source files. If an independent fresh receiver is available and its use is authorized, let it attempt the next action using only the handoff and accessible inputs, and record what it actually did. Otherwise label the check “Source check only; fresh receiver not tested.” Do not simulate a receiver and call that independent proof. Do not send the handoff to another chat/person, run the next task or grant access unless separately authorized. Return the saved handoff and remaining access gaps.

## Required inputs

- Authorized project files and current work record
- Scope, desired outcome and the receiver’s known access

## Expected outputs

- Self-contained handoff file
- Source and validation references
- Next action with stop conditions

## Checks

- The next action is possible from the supplied handoff and accessible files.
- Actual validation, unrun checks and open decisions are distinct.
- Private data and inaccessible machine paths are not mistaken for usable shared inputs.
- A source check is not labeled an independent receiver test.

## Limits

- Separate chats do not automatically share context, files or permissions.
- A handoff cannot transfer credentials or computer access by describing them.
- Creating a new chat, messaging it and executing its next task require actual authorization.

Revision: 2.0.1. [Read the complete kit](files/README.md) before use. This optional format is not a claim of support by every tool.
Open SKILL.md ↗