Manus Monitor
AdvisoryAudited by Static analysis on Apr 30, 2026.
Overview
No suspicious patterns detected.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Prompts and requested work may be performed by Manus outside OpenClaw, using Manus's own web/tool capabilities.
The skill delegates user prompts to an external autonomous agent, which is central to the stated purpose but important for users to understand.
Use the Manus API to create autonomous AI tasks. Manus can browse the web, use tools, and deliver complete results
Use this skill only for tasks you are comfortable sending to Manus, and avoid including secrets or highly sensitive data in prompts unless Manus is trusted for that data.
Anyone running the skill with this key may be able to act on the connected Manus account within the API key's permissions.
The skill requires a Manus API key to create, list, and retrieve tasks. This is expected for the integration and no artifact shows the key being sent anywhere other than Manus.
Header: `API_KEY: <your-key>` Set via: - `MANUS_API_KEY` env var
Use a dedicated or scoped Manus API key if available, rotate it if exposed, and avoid sharing the environment with untrusted code.
It is harder to verify that the reviewed package identity matches the registry listing the user expects.
The internal package metadata differs from the registry information supplied for this review, which lists a different owner ID, slug, and version.
"ownerId": "kn7d7xy7794nh6aaabfga5wwzh7zptdm", "slug": "manus", "version": "1.1.0"
Confirm the publisher and package provenance before installing, and align registry metadata with the packaged _meta.json.
Private task names, descriptions, statuses, and links could be sent to a Telegram chat without the user realizing this skill uses Telegram.
The monitor can send task status, task description/title, and a Manus task link to Telegram when TELEGRAM_BOT_TOKEN and TELEGRAM_CHAT_ID are present, but SKILL.md and the requirements do not disclose this Telegram data flow.
local message="Manus 任务状态更新 ... 链接：https://manus.im/app/${task_id}" ... curl -s -X POST "https://api.telegram.org/bot${bot_token}/sendMessage"Disclose Telegram notifications in SKILL.md, declare the Telegram environment variables, require explicit opt-in, and avoid sending sensitive task details or links by default.
The skill may continue running periodic agent activity and API checks after the initial task, with little visible delivery to the user.
The script creates a recurring background OpenClaw agent-turn cron job every minute. This persistent monitoring behavior is not described in SKILL.md.
"name": "manus-task-monitor", "schedule": { "kind": "every", "everyMs": 60000 }, "payload": { "kind": "agentTurn", "message": "执行 $MANUS_SCRIPT 脚本检查 Manus 任务状态..." }, "delivery": { "mode": "none" }, "enabled": trueRequire explicit user approval before creating the cron job, document how to stop it, make monitoring opt-in, and ensure the job reliably removes itself when no tasks remain.