openclone

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow includes /openclone ingest <URL> (and YouTube transcript ingestion), which fetches outsider-authored public web content and stores/feeds its readable text into the clone’s local markdown context for future LLM conversations.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's install and plugin commands (git clone ... https://github.com/open-clone/openclone && ./setup, /plugin marketplace add https://github.com/open-clone/openclone and the skills add -g https://github.com/akillness/oh-my-gods ...) fetch remote repositories and run setup/install steps, which downloads and executes remote code at install/runtime.