grok-search

**Verdict:** A strong, agent-friendly Grok-based search skill, particularly valuable when you need X/Twitter content or structured JSON with citations. Best used as a complementary search source rather than a sole oracle. **What it does well** - Wraps xAI Grok’s `web_search` and `x_search` tools in Node scripts, returning clean JSON with citations. - Handles both web and X/Twitter search, including useful X filters (date range, handles include/exclude, result limits). - Output schema is stable and agent-oriented, making it easy to plug into automated pipelines. **Risks / limitations** - Hard dependency on xAI’s API (network, uptime, rate limits, and valid `XAI_API_KEY`). - No explicit safety or content filtering; X/Twitter and web results can include NSFW, biased, or misleading content. - Search quality and recall are bounded by Grok’s underlying tools and may differ from mainstream engines. **Recommended scenarios** - Workflows that need tweets/threads or handle-targeted X/Twitter monitoring with structured output. - Agent chains where reliable citations and machine-readable search results are more important than raw search breadth. - Situations where you want Grok’s perspective as an additional search channel alongside other providers.