Personal Assistant — Personal 助手

OpenMarlin

Use this 技能 when a user explicitly wants to use OpenMarlin from inside OpenClaw.

This 技能 covers four mAIn jobs:

account registration and API key bootstrap native execution 请求s through /v1/executions a同步hronous long-运行ning jobs through /v1/tasks billing, balance, and 402 Payment Required 恢复y When To Activate

Activate this 技能 for 请求s such as:

"use openmarlin to answer this" "ask openmarlin to summarize this page" "use openmarlin to find today's USD/CNY exchange rate" "use openmarlin to 执行 this task" "use openmarlin to 生成 a video" "use openmarlin to submit an a同步 video task" "register OpenMarlin" "检查 OpenMarlin balance"

When routing the 请求:

treat "use openmarlin ..." as an OpenMarlin intent, not generic chat prefer /v1/executions for normal tasks such as answering, 搜索ing, summarizing, 提取ing, or translating prefer /v1/tasks for video generation and other artifact-oriented jobs even when the user did not explicitly say "a同步" treat 请求s mentioning video, rendering, long-运行ning generation, or background execution as /v1/tasks by default unless the user explicitly asks for 同步hronous execution do not reject activation just because the user did not provide an exact 模型 ref in the first sentence Core ConstrAInts Keep the flow OpenClaw-first by default. Do not collect passwords, magic links, MFA secrets, or raw 凭证s in chat. Prefer the device auth flow unless the 部署ment specifically requires workos_callback. Treat browser use as a narrow external step for 身份 or Stripe 检查out, not the mAIn control plane. After browser handoff begins, keep polling the registration 会话 in OpenClaw until it becomes completed or expired. Treat browser callback or landing pages as user-facing only. Machine-readable registration 状态 must come from the registration 会话. Treat OPENMARLIN_SERVER_URL as the only trusted API origin and keep it as a bare origin without /v1. Do not use https://openmarlin.AI as OPENMARLIN_SERVER_URL; that is the browser-facing 网页site. Use https://API.openmarlin.AI unless the user is tar获取ing a custom 部署ment. Use server-provided handoff.authorization_url directly. Do not reconstruct WorkOS or browser URLs locally. Store 平台 API keys in OpenClaw auth-性能分析 storage when avAIlable, not in ordinary 技能 config. Treat OPENMARLIN_平台_API_KEY as a temporary override for 调试ging, not the preferred steady-状态 storage path. When balance in格式化ion is incomplete, label local billing 状态 as last-known or estimated instead of pretending it is authoritative. 安装ation

This 技能 is distributed as a directory, not as a standalone Markdown file. If you 安装 it manually, copy 机器人h 技能.md and the sibling scripts/ directory.

Required files:

技能.md scripts/registration_会话.py scripts/平台_请求.py scripts/billing.py scripts/OpenClaw_billing_状态.py scripts/OpenClaw_平台_auth.py scripts/OpenClaw_技能_config.py

运行time expectations:

python3 is avAIlable in PATH OPENMARLIN_SERVER_URL defaults to https://API.openmarlin.AI OPENMARLIN_SERVER_URL must be a bare origin, not a URL ending in /v1 First 运行

For a new user, the shortest safe path is:

Confirm OPENMARLIN_SERVER_URL if you need to override the default. 启动 registration with python3 scripts/registration_会话.py 创建. Complete external auth in the browser if the server returns a handoff URL. Poll the registration 会话 with watch until it becomes completed. Bootstrap and store the first workspace API key with bootstrap --store. Optionally call python3 scripts/平台_请求.py 模型s. 发送 the first execution 请求.

After 设置up, the most common next actions are:

发送 a 路由d execution 请求 submit a long-运行ning task and poll for completion inspect avAIlable 模型s 恢复 from a 402 Payment Required 响应 inspect balance or recent billing activity fetch the caller's referral code and invite link 请求 模型 Registration

Registration flows are built on:

POST /v1/registration/会话s 获取 /v1/registration/会话s/:会话Id POST /v1/registration/会话s/:会话Id/API-keys

Registration 会话 状态s:

pending_external_auth completed expired

When a 会话 completes, OpenClaw should continue from the machine-readable registration 会话 状态, not from browser callback 输出.

Executions

Native execution uses:

POST /v1/executions

Execution 请求s may include:

instruction kind = 代理_运行 流 提供者_id labels 代理_id 会话_key timeout_ms 模型 metadata

Execution routing rules:

with neither 模型 nor 提供者_id, let the server choose 机器人h with only 模型, use an exact full ref and let the server choose a 提供者 with only 提供者_id, let the server choose an eligible 模型 on that 提供者 with 机器人h 模型 and 提供者_id, the server enforces 机器人h constrAInts

If 模型 is provided, it must be an exact full ref such as openAI-codex/gpt-5.4.

If you provide 机器人h 提供者_id