x-search

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

• Third-party content exposure detected (medium risk: 0.30). The runtime path uses scripts/search-x.mjs to run Grok with a user-supplied --query prompt, obtains Grok’s search results, extracts https://x.com/... URLs, and then (per validation rules) the skill relies on those posts’ content read at runtime (via the logged-in browsing step described in SKILL.md).

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

• Potentially malicious external URL detected (high risk: 0.90). The script sets GROK_CLI_CHAT_PROXY_BASE_URL to https://cli-chat-proxy.grok.com/v1 and invokes the local Grok CLI at runtime, which will contact that proxy to obtain model responses that directly determine the skill's returned answer and citations.