agent-export

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Agent Export — Migration Bundle Creator

Agent导出——迁移包创建工具

Create a structured migration bundle from your current agent data and upload it to the Starchild migration relay. The receiving Starchild agent uses the

agent-import

skill to load it.

从当前Agent数据创建结构化迁移包，并上传至Starchild迁移中继服务。接收方Starchild Agent将使用

agent-import

技能加载该包。

Migration Bundle Format

迁移包格式

The bundle is a tar.gz archive with this structure:

migration/
  manifest.json           # required — metadata
  memory/
    agent.json            # agent's own notes & knowledge
    user.json             # what the agent knows about the user
  identity/
    profile.json          # agent name, personality
    soul.md               # behavioral guidelines (free-form markdown)
  user/
    settings.json         # user preferences (name, timezone, language)
  tasks/
    tasks.json            # scheduled/recurring tasks
  env/
    keys.json             # environment variable names needed (NO values)
  files/                  # arbitrary files to carry over
    ...

All files are optional except

manifest.json

该包是一个tar.gz归档文件，结构如下：

migration/
  manifest.json           # 必填——元数据
  memory/
    agent.json            # Agent自身的笔记与知识
    user.json             # Agent对用户的认知
  identity/
    profile.json          # Agent名称、性格
    soul.md               # 行为准则（自由格式Markdown）
  user/
    settings.json         # 用户偏好（姓名、时区、语言）
  tasks/
    tasks.json            # 计划/定期任务
  env/
    keys.json             # 所需环境变量名称（不含值）
  files/                  # 需迁移的任意文件
    ...

除

manifest.json

外，所有文件均为可选。

File Specifications

文件规范

manifest.json (required)

manifest.json（必填）

json

{
  "version": "1.0",
  "source": "openclaw",
  "created_at": "2025-07-13T10:00:00Z",
  "description": "Migration from OpenClaw agent"
}

```
source
```
: identifier of the originating agent/platform (free text)
```
version
```
: always
```
"1.0"
```
for now

json

{
  "version": "1.0",
  "source": "openclaw",
  "created_at": "2025-07-13T10:00:00Z",
  "description": "Migration from OpenClaw agent"
}

```
source
```
: 源Agent/平台的标识符（自由文本）
```
version
```
: 目前固定为
```
"1.0"
```

memory/agent.json

Agent's accumulated knowledge — things the agent learned about the environment, tool quirks, API notes, workflows, conventions.

json

{
  "entries": [
    "Coinglass funding rate values are already in percent; do not multiply by 100.",
    "User's Hyperliquid account uses cross-margin by default.",
    "For Fly.io deploys, extract FLY_TOKEN via sed from .env in project dir."
  ]
}

Each entry: 1-3 sentences, one concern per entry. Think "what would I need to remember next session?"

Agent积累的知识——关于环境、工具特性、API笔记、工作流、惯例等内容。

json

{
  "entries": [
    "Coinglass funding rate values are already in percent; do not multiply by 100.",
    "User's Hyperliquid account uses cross-margin by default.",
    "For Fly.io deploys, extract FLY_TOKEN via sed from .env in project dir."
  ]
}

每条条目：1-3句话，每条对应一个关注点。思考“下一次会话我需要记住什么？”

memory/user.json

What the agent knows about the user — their role, preferences, communication style, interests.

json

{
  "entries": [
    "Prefers concise responses under 25 lines, direct conclusions, no hedges.",
    "Technical background in full-stack dev and crypto trading.",
    "Located in Argentina, primary language is Chinese."
  ]
}

Agent对用户的认知——用户的角色、偏好、沟通风格、兴趣。

json

{
  "entries": [
    "Prefers concise responses under 25 lines, direct conclusions, no hedges.",
    "Technical background in full-stack dev and crypto trading.",
    "Located in Argentina, primary language is Chinese."
  ]
}

identity/profile.json

json

{
  "name": "MyAgent",
  "vibe": "professional, concise, opinionated",
  "emoji": "🤖",
  "creature": "robot"
}

All fields optional.

vibe

is a short personality description.

json

{
  "name": "MyAgent",
  "vibe": "professional, concise, opinionated",
  "emoji": "🤖",
  "creature": "robot"
}

所有字段均为可选。

vibe

是简短的性格描述。

identity/soul.md

Free-form markdown describing how the agent should behave. Keep it under 50 lines. Example:

markdown

undefined

自由格式Markdown，描述Agent的行为方式。控制在50行以内。示例：

markdown

undefined

Behavior

Be concise, skip filler phrases
Have opinions, back them with data
For trading: present analysis, not financial advice

undefined

Be concise, skip filler phrases
Have opinions, back them with data
For trading: present analysis, not financial advice

undefined

user/settings.json

json

{
  "name": "Alice",
  "what_to_call": "Boss",
  "timezone": "Asia/Shanghai",
  "language": "zh-CN"
}

```
timezone
```
: IANA format (e.g.,
```
America/New_York
```
,
```
Asia/Tokyo
```
)
```
language
```
: BCP-47 code (
```
en
```
,
```
zh-CN
```
,
```
ja
```
, etc.)

json

{
  "name": "Alice",
  "what_to_call": "Boss",
  "timezone": "Asia/Shanghai",
  "language": "zh-CN"
}

```
timezone
```
: IANA格式（例如
```
America/New_York
```
、
```
Asia/Tokyo
```
）
```
language
```
: BCP-47代码（
```
en
```
、
```
zh-CN
```
、
```
ja
```
等）

tasks/tasks.json

json

{
  "tasks": [
    {
      "title": "BTC Price Alert",
      "schedule": "every 30 minutes",
      "description": "Check BTC price, alert if > $100k or < $80k",
      "channels": ["web", "tg"]
    },
    {
      "title": "Daily Market Summary",
      "schedule": "0 1 * * *",
      "description": "Summarize crypto market at 9am Shanghai time"
    }
  ]
}

```
schedule
```
: cron expression (UTC), interval (
```
every 5 minutes
```
), or delay (
```
in 2 hours
```
)
```
channels
```
: optional, subset of
```
["web", "tg", "wechat"]
```

json

{
  "tasks": [
    {
      "title": "BTC Price Alert",
      "schedule": "every 30 minutes",
      "description": "Check BTC price, alert if > $100k or < $80k",
      "channels": ["web", "tg"]
    },
    {
      "title": "Daily Market Summary",
      "schedule": "0 1 * * *",
      "description": "Summarize crypto market at 9am Shanghai time"
    }
  ]
}

```
schedule
```
: cron表达式（UTC）、时间间隔（
```
every 5 minutes
```
）或延迟时间（
```
in 2 hours
```
）
```
channels
```
: 可选，为
```
["web", "tg", "wechat"]
```
的子集

env/keys.json

List environment variable names the agent needs (values are entered separately for security).

json

{
  "keys": [
    {"key": "OPENAI_API_KEY", "label": "OpenAI API Key", "required": true},
    {"key": "TELEGRAM_BOT_TOKEN", "label": "Telegram Bot Token", "required": false}
  ]
}

列出Agent所需的环境变量名称（值需单独输入以保障安全）。

json

{
  "keys": [
    {"key": "OPENAI_API_KEY", "label": "OpenAI API Key", "required": true},
    {"key": "TELEGRAM_BOT_TOKEN", "label": "Telegram Bot Token", "required": false}
  ]
}

files/

Put any arbitrary files here. They'll be copied to the Starchild workspace as-is, preserving subdirectory structure. Use for custom scripts, configs, dashboards, etc.

将任意文件放入此处。它们将原样复制到Starchild工作区，保留子目录结构。可用于自定义脚本、配置文件、仪表板等。

Export Workflow

导出工作流

Step 1 — Gather Data

步骤1 — 收集数据

Collect information from your current agent. Map your data to the formats above:

Chat history / memory → extract key facts into
```
memory/agent.json
```
and
```
memory/user.json
```
System prompt / persona → transform into
```
identity/profile.json
```
+
```
identity/soul.md
```
User preferences →
```
user/settings.json
```
Recurring automations →
```
tasks/tasks.json
```
API keys needed →
```
env/keys.json
```
(names only!)
Custom files → copy into
```
files/
```

从当前Agent收集信息，并将数据映射为上述格式：

聊天记录/记忆 → 提取关键事实至
```
memory/agent.json
```
和
```
memory/user.json
```
系统提示词/角色设定 → 转换为
```
identity/profile.json
```
+
```
identity/soul.md
```
用户偏好 →
```
user/settings.json
```
定期自动化任务 →
```
tasks/tasks.json
```
所需API密钥 →
```
env/keys.json
```
（仅名称！）
自定义文件 → 复制至
```
files/
```

Step 2 — Build the Bundle

步骤2 — 构建迁移包

bash

undefined

bash

undefined

Create directory structure

mkdir -p migration/{memory,identity,user,tasks,env,files}

Write your JSON files (examples — replace with real data)

cat > migration/manifest.json << 'EOF' {"version": "1.0", "source": "my-agent", "created_at": "2025-07-13T10:00:00Z"} EOF

... write other files ...

Pack into tar.gz

cd migration && tar czf ../migration-bundle.tar.gz . && cd ..

undefined

cd migration && tar czf ../migration-bundle.tar.gz . && cd ..

undefined

Step 3 — Upload to Relay

步骤3 — 上传至中继服务

bash

RESPONSE=$(curl -s -X POST https://sc-agent-migration.fly.dev/paste \
  -H "Content-Type: application/octet-stream" \
  --data-binary @migration-bundle.tar.gz)

CODE=$(echo "$RESPONSE" | grep -o '"code":"[^"]*"' | cut -d'"' -f4)
TOKEN=$(echo "$RESPONSE" | grep -o '"download_token":"[^"]*"' | cut -d'"' -f4)

echo "Migration code:   $CODE"
echo "Download token:   $TOKEN"

Give both values to the user — they need the code AND the token to import in Starchild.

⚠️ The code expires in 1 hour and is single-use (deleted after first successful download). ⚠️ The token is the only authorization for download — do not lose it.

bash

RESPONSE=$(curl -s -X POST https://sc-agent-migration.fly.dev/paste \
  -H "Content-Type: application/octet-stream" \
  --data-binary @migration-bundle.tar.gz)

CODE=$(echo "$RESPONSE" | grep -o '"code":"[^"]*"' | cut -d'"' -f4)
TOKEN=$(echo "$RESPONSE" | grep -o '"download_token":"[^"]*"' | cut -d'"' -f4)

echo "Migration code:   $CODE"
echo "Download token:   $TOKEN"

将两个值都提供给用户——他们需要代码和令牌才能在Starchild中完成导入。

⚠️ 代码1小时后过期且仅限单次使用（首次成功下载后即被删除）。 ⚠️ 令牌是下载的唯一授权凭证——请勿丢失。

Tips for Other Agents

其他Agent的注意事项

Don't overthink it. Even just
```
manifest.json
```
+
```
memory/user.json
```
with 5 entries is useful.
Summarize, don't dump. Convert raw chat history into concise memory entries.
Timezone matters. Get the user's timezone right — it affects all scheduled tasks.
No secrets in the bundle. API keys go in
```
env/keys.json
```
as names only.
Test your tar.gz —
```
tar tzf migration-bundle.tar.gz
```
should show paths starting with
```
./
```
.

无需过度纠结。即使只有
```
manifest.json
```
+
```
memory/user.json
```
且包含5条条目，也会很有用。
总结而非转存。将原始聊天记录转换为简洁的记忆条目。
时区很重要。准确获取用户时区——这会影响所有计划任务。
迁移包中不要包含机密信息。API密钥仅在
```
env/keys.json
```
中填写名称。
测试你的tar.gz文件——执行
```
tar tzf migration-bundle.tar.gz
```
应显示以
```
./
```
开头的路径。