Loading ToolBench
Exposes Grok Build's live X/web search as a Model Context Protocol tool, enabling any MCP-capable coding agent (Claude Code, Codex, Cursor) to call it autonomously for real-time information retrieval.
Single tool 'grok_search' has a well-crafted description (189 chars) that clearly states WHAT (search X/web in real-time), WHEN (current info beyond training cutoff), and WHAT IT RETURNS (synthesized answer with URLs). Input schema is present with proper type and description. However, the schema lacks enum constraints for the query parameter, and no output schema is documented. The tool is read-only and properly marked. Naming follows verb_noun convention. Overall solid for a single-tool server, but output documentation and parameter constraints would elevate it.
Search X (Twitter) and the web in REAL TIME using Grok, and return a synthesized answer with source URLs. Use this whenever you need current or recent information that may be beyond your training cutoff: latest package/library versions, breaking API changes, recent releases, ongoing incidents, or what people are saying on X right now. Read-only — it never edits files or runs shell commands. Prefer this over guessing about current events, social sentiment, or package versions.
Output schema not documented. LLMs cannot plan downstream operations or extract structured data without knowing what fields grok_search returns.
Query parameter lacks format constraints or examples. Description says 'Be specific; mention X/x.com if you want social or primary sources' but no enum, pattern, or length bounds are declared in the schema.
No error handling guidance in tool description. What happens if the query is malformed, the Grok CLI is unavailable, or rate limits are hit? LLMs need recovery hints.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 64 | <=2025-11-25 | v2 |