Agent Skillsletta-ai/letta-code › teleporting-between-environments

teleporting-between-environments

GitHub

实现 Agent 在不同计算环境(云端、本地等)间的无缝切换,确保会话记忆延续并协调文件传输与环境配置准备。

src/skills/builtin/teleporting-between-environments/SKILL.md letta-ai/letta-code

Trigger Scenarios

用户要求将任务转移到云端或另一台计算机继续 需要在环境切换前上传或下载工件 明确提到在环境间 'teleport' 或移动工作

Install

npx skills add letta-ai/letta-code --skill teleporting-between-environments -g -y
More Options

Non-standard path

npx skills add https://github.com/letta-ai/letta-code/tree/main/src/skills/builtin/teleporting-between-environments -g -y

Use without installing

npx skills use letta-ai/letta-code@teleporting-between-environments

指定 Agent (Claude Code)

npx skills add letta-ai/letta-code --skill teleporting-between-environments -a claude-code -g -y

安装 repo 全部 skill

npx skills add letta-ai/letta-code --all -g -y

预览 repo 内 skill

npx skills add letta-ai/letta-code --list

SKILL.md

Frontmatter
{
    "name": "teleporting-between-environments",
    "description": "Moves the current agent conversation to Cloud, Desktop Local, or another Cloud-registered environment while coordinating machine-local files and setup. Use when the user says \"let's continue this task on cloud\", asks to continue or move work on another connected computer, wants to teleport between environments, or needs to upload or download artifacts before a handoff."
}

Teleporting Between Environments

Teleport the current agent and conversation without losing conversational memory. Prepare machine-local state first, then let the destination continue the same task.

Mental model

  • Memory follows the agent; filesystem access does not. Files, working directories, credentials, running processes, and local services belong to the computer currently executing the conversation.
  • Upload and download paths are relative to the current computer. Cloud cannot read a laptop path until the conversation teleports to that laptop.
  • The conversation’s managed Cloud sandbox remains alive while the conversation runs elsewhere.
  • Filesystem paths and cwd do not transfer between computers. Re-establish the destination’s repository, working directory, dependencies, credentials, and services after arrival.

Commands

letta teleport list
letta teleport cloud
letta teleport local
letta teleport <environment>
  • list: show accessible online Cloud-registered targets.
  • cloud: create or resume this conversation’s managed Cloud sandbox.
  • local: target the one online Letta Desktop environment. Desktop Remote Access must be enabled. If several Desktop environments are online, use list and target one explicitly.
  • <environment>: target a listed remote environment by its friendly selector.

Do not run or recommend letta teleport back; return to Desktop with local.

Transfer files with the existing sandbox commands:

letta sandbox upload <local-path>
letta sandbox download <sandbox-path> [--to <local-path>]

Do not invent letta teleport push, pull, or remote file-listing commands. No sandbox ID or wake command is needed.

Prepare the handoff

Before teleporting:

  1. Identify the target and every file, repository state, credential, service, or setup step the task needs there.
  2. Finish work that requires the current computer. Verify relevant files exist before transferring them.
  3. Upload current-computer artifacts needed in Cloud and retain the returned /root/downloads/... paths in context.
  4. Retain enough context to recreate destination-local setup, including the repository, branch or revision, working directory, dependency commands, and next action.
  5. Run teleport only after all source-side preparation is complete.

Teleport is the final action

Run the teleport command as the only command in the final Bash tool call. Do not chain later commands, poll the teleport operation, or invoke another source-side tool after it.

The CLI intentionally returns after the server accepts the handoff. Once the Bash result is persisted, the source yields at a clean turn or tool boundary and the destination resumes with no synthetic user message.

If the command reports an offline, stale, unsupported, same-source, or startup error, the conversation remains on the source. Surface the concrete error, correct it if possible, and retry only after the target is available.

Common workflows

Continue local work in Cloud

  1. Inspect the local task state and identify local-only artifacts or setup.

  2. Upload each artifact Cloud needs:

    letta sandbox upload <local-path>
    
  3. Retain each returned sandbox path.

  4. As the final action, run:

    letta teleport cloud
    
  5. After continuation in Cloud, establish the Cloud-local cwd and repository setup before using the uploaded paths.

Return to Desktop Local

  1. Retain the next local action and any setup the Desktop environment needs.

  2. Confirm Desktop is open with Remote Access enabled. If more than one Desktop environment is online, use letta teleport list and choose one explicitly.

  3. As the final action, run:

    letta teleport local
    

Run a separate headless turn on an environment

Use --environment when a separate headless invocation, rather than the current conversation handoff, should execute on Cloud or another online environment:

letta -p --conversation <id> --environment cloud "<prompt>"
letta -p --conversation <id> --environment <name|device-id|connection-id> "<prompt>"

This routes that headless message only. Use letta teleport ... to move the currently executing conversation.

Continue on another connected computer

  1. Discover available targets if needed:

    letta teleport list
    
  2. Prepare or upload everything the current computer owns.

  3. As the final action, run:

    letta teleport <environment>
    

Version History

  • ee230f3 Current 2026-08-19 17:32

    新增支持通过CLI teleport local命令直接切换至桌面本地环境,此前仅支持云端及远程目标。

  • c0a387c 2026-08-13 16:15

Same Skill Collection

.skills/adding-models/SKILL.md
.skills/capturing-tui-visual-proof/SKILL.md
src/skills/builtin/acquiring-skills/SKILL.md
src/skills/builtin/context-doctor/SKILL.md
src/skills/builtin/converting-mcps-to-skills/SKILL.md
src/skills/builtin/creating-mods/SKILL.md
src/skills/builtin/creating-skills/SKILL.md
src/skills/builtin/customizing-commands/SKILL.md
src/skills/builtin/customizing-statusline/SKILL.md
src/skills/builtin/dispatching-coding-agents/SKILL.md
src/skills/builtin/editing-letta-code-desktop-preferences/SKILL.md
src/skills/builtin/finding-agents/SKILL.md
src/skills/builtin/generating-mod-envs/SKILL.md
src/skills/builtin/image-generation/SKILL.md
src/skills/builtin/initializing-memory/SKILL.md
src/skills/builtin/letta-guide/SKILL.md
src/skills/builtin/managing-shared-memory/SKILL.md
src/skills/builtin/messaging-agents/SKILL.md
src/skills/builtin/migrating-memory/SKILL.md
src/skills/builtin/modifying-the-harness/SKILL.md
src/skills/builtin/scheduling-tasks/SKILL.md
src/skills/builtin/self-configuration/SKILL.md
src/skills/builtin/syncing-memory-filesystem/SKILL.md
src/skills/builtin/using-cloud-mcp/SKILL.md

Metadata

Files
0
Version
0016701
Hash
414e3c7c
Indexed
2026-08-13 16:15

trang chủ - Wiki
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-08-29 21:32
浙ICP备14020137号-1 $bản đồ khách truy cập$