100Critical
The skill is vulnerable to prompt injection and persistently stores sensitive user credentials locally, risking exfiltration by other processes.
Behavior Analysis
Description verified — behavior matches claim
Claims to do
淘宝图片搜索技能: - 优先执行脚本：`run-taobao-task.js`。 - 脚本失败、登录超时或页面结构变化时，回退 `browser` 工具。 - 默认不下单、不支付；仅搜索与加购。
Actually does
This skill installs Playwright and Chromium, then executes a Node.js script (`run-taobao-task.js`) to automate image search and add-to-cart actions on `https://www.taobao.com`. It handles automatic login, uploads images, navigates product pages, and adds items to the cart, persisting browser sessions (including sensitive login tokens) locally. If the script fails, it falls back to a `browser` tool for manual interaction.
Skill Info
Namelazygunner/taobao-image-search
Registryclawhub
Version1.1.3
PURLpkg:clawhub/lazygunner/taobao-image-search@1.1.3
Stars4
Downloads1,423
Installs4
Source
SHA-256326ef76e6b14...
SHA-18cf3417cca6b...
MD56ca081a914ed...
TLSH1e81d92e8546...
Size3,870 bytes
Install
This skill has been flagged as potentially malicious. Review the findings below before installing.
ClawHubDocs
openclaw skills install lazygunner/taobao-image-searchAssessments (2)
AI Agent Traps ↗Prompt Injection Detectedcriticalprompt injection
DeBERTa classifier detected prompt injection (confidence: 0.96)
96% confidence
Storage of Sensitive User Credentialshighhardcoded secrets
The skill explicitly stores sensitive user credentials (Taobao login tokens, cookies, storage state) in local files (`taobao-storage-state.json`, `.pw-user-data-taobao/`). While a warning is provided, this creates a persistent artifact that could be targeted by other malicious processes or exfiltrated if the skill's script were compromised.
100% confidenceline 68
这些文件包含您的淘宝登录令牌，请务必将其视为敏感凭据，严禁上传或分享。
Badge
Markdown
[](https://mondoo.com/ai-agent-security/skills/clawhub/lazygunner/taobao-image-search)HTML
<a href="https://mondoo.com/ai-agent-security/skills/clawhub/lazygunner/taobao-image-search"><img src="https://mondoo.com/ai-agent-security/api/badge/clawhub/lazygunner/taobao-image-search.svg" alt="Mondoo Skill Check" /></a>Image URL
https://mondoo.com/ai-agent-security/api/badge/clawhub/lazygunner/taobao-image-search.svgSecure your AI agents
Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.
- Continuous scanning of skills across all registries
- Policy enforcement before skills reach your agents
- Integration with your existing security workflow