nl2ms-ui

PassAudited by ClawScan on May 10, 2026.

Overview

This instruction-only skill coherently generates Midscene automation scripts, but users should review generated scripts because examples can automate account/device actions and reuse saved credentials or login state.

Use this skill as a script generator, not as a blindly trusted executor. Before running generated Midscene scripts, check the target app or URL, account/session being used, message or form content, and any permission or agreement-handling steps.

Findings (3)

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.

Low

#ASI02: Tool Misuse and Exploitation

What this means

A generated script could click through dialogs, grant permissions, or send messages in an app if the user executes it unchanged.

Why it was flagged

The example automation can accept permissions/agreements and send a message if the generated script is later run. This is consistent with UI test automation, but it is a state-changing capability users should review.

Skill content

aiActionContext: "处理弹窗和权限请求。如果出现位置权限、用户协议等弹窗，点击同意。..."
...
- aiTap: 发送

Recommendation

Review generated steps before running them, use test accounts/devices where possible, and avoid auto-accepting permissions or agreements unless that is intentional.

Low

#ASI03: Identity and Privilege Abuse

What this means

If run with a real saved session, generated web automation may act as the logged-in user.

Why it was flagged

The example shows using a saved Playwright login state. This is normal for authenticated browser testing and there is no evidence of leakage, but it means generated scripts may act with the user's saved session privileges.

Skill content

if (fs.existsSync(storageStatePath)) {
  console.log("使用存储的登录状态...");
  contextOptions.storageState = storageStatePath;
}

Recommendation

Keep storage state files private, prefer test accounts, and confirm the target URL and actions before executing generated scripts.

Info

#ASI03: Identity and Privilege Abuse

What this means

Generated projects may depend on local API keys, so accidental sharing of .env files could expose credentials.

Why it was flagged

The skill advises managing API keys through .env files. This is expected for Midscene/API-based automation and no artifact shows hardcoded secrets or credential exfiltration.

Skill content

使用 `.env` 管理 API 密钥，设置合适的 `aiActionContext`

Recommendation

Do not commit or share .env files, and use least-privilege API keys for automation.