This skill is vulnerable to prompt
Claims to do
agent-im 会话(生图 / 生视频): 通过 agent-im 的 OpenAPI 创建会话、发送消息(生图、生视频、编辑视频等)、上传图片/视频文件,并查询会话消息进展。
Actually does
This skill executes Python scripts to interact with the LibTV OpenAPI. It creates and manages sessions, sends messages for AI image/video generation and editing, uploads local image/video files (up to 200MB) to an OSS, queries session progress, and downloads generated media from specified URLs. All API calls are authenticated using a `LIBTV_ACCESS_KEY`.
openclaw skills install 316530790/libtv-skillsDeBERTa classifier detected prompt injection (confidence: 1.00)
The skill explicitly executes Python scripts (`python3 {baseDir}/scripts/*.py`) where arguments can be influenced by user input. Specifically, `upload_file.py` takes a local file path, and `download_results.py` can download from arbitrary URLs (`--urls`) and write to arbitrary local paths (`--output-dir`). This capability enables potential arbitrary file uploads (data exfiltration), downloads of malicious content from attacker-controlled URLs, and writing to sensitive file system locations (system compromise, resource abuse).
python3 {baseDir}/scripts/upload_file.py /path/to/image.png
python3 {baseDir}/scripts/download_results.py SESSION_ID --output-dir ~/Desktop/my_project
python3 {baseDir}/scripts/download_results.py --urls URL1 URL2 URL3[](https://mondoo.com/ai-agent-security/skills/clawhub/316530790/libtv-skills)<a href="https://mondoo.com/ai-agent-security/skills/clawhub/316530790/libtv-skills"><img src="https://mondoo.com/ai-agent-security/api/badge/clawhub/316530790/libtv-skills.svg" alt="Mondoo Skill Check" /></a>https://mondoo.com/ai-agent-security/api/badge/clawhub/316530790/libtv-skills.svgSkills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.