AIEO monitoring

ReviewAudited by ClawScan on May 10, 2026.

Overview

This instruction-only skill appears aligned with AIEO monitoring, but it may automate logged-in AI services, send brand test questions to third-party platforms, and save reports/screenshots locally.

Before installing, confirm the AI platforms and question set, avoid including confidential strategy or customer data in prompts, use dedicated test accounts where possible, and keep the generated monitoring/ reports and screenshots private.

Findings (5)

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.

What this means

The agent may drive browser sessions and, if it chose to use Bash, could run local commands, although no unsafe shell commands are shown in the artifacts.

Why it was flagged

The skill can automate browser interactions with AI platforms and also has Bash available. Browser automation is central to the monitoring purpose; Bash is broader local authority that users should notice.

Skill content
allowed-tools: Read, Write, Bash, Grep, Glob, WebFetch, WebSearch, mcp__playwright__browser_navigate, ... mcp__playwright__browser_type, ... mcp__playwright__browser_take_screenshot
Recommendation

Run it only for intended monitoring tasks, review the platforms/questions first, and consider removing or avoiding Bash access if shell commands are not needed.

What this means

Test questions may be associated with the user's AI platform account history or usage limits.

Why it was flagged

Several monitored AI platforms may require authenticated sessions, so the automation may operate under the user's platform accounts.

Skill content
| ChatGPT | chat.openai.com | 需登录 | ... | Claude | claude.ai | 需登录 | ... | 豆包 | doubao.com | 需登录 |
Recommendation

Use dedicated test accounts where possible and avoid running the monitoring through personal accounts unless you accept the account-history and usage implications.

What this means

If granted access, the agent could view business performance data or customer-system metrics needed for the report.

Why it was flagged

The business-metrics portion may rely on privileged analytics, ecommerce, CRM, or advertising accounts. This is purpose-aligned for conversion tracking but sensitive.

Skill content
数据来源: - GA4(Google Analytics 4) - 百度统计 ... 数据来源: - 电商后台(天猫、京东) - CRM 系统 - 广告后台
Recommendation

Provide only the minimum metrics needed, avoid exposing customer-level data, and confirm any analytics/CRM account access before use.

What this means

Brand questions, positioning details, and test prompts may leave the local workspace and be processed by third-party AI services.

Why it was flagged

The workflow sends monitoring questions to external AI platforms and records their answers. This external data flow is expected for AIEO monitoring but can expose sensitive prompts or brand strategy.

Skill content
依次访问各平台 ... 对每个问题: a. 输入问题 ... d. 记录回答内容
Recommendation

Review the question library for confidential information and only submit data you are comfortable sharing with the selected AI platforms.

What this means

Saved reports and screenshots may retain brand strategy, AI responses, and business metrics across future sessions.

Why it was flagged

The skill intentionally creates persistent local reports and screenshots containing AI answers and monitoring evidence.

Skill content
监控报告保存至 monitoring/ 目录 ... 截图保存 ... monitoring/screenshots/
Recommendation

Store the monitoring/ directory securely, review generated files before sharing, and delete old reports when no longer needed.