Ask Grok or Gemini via opencli (browser-session based, no API cost). Use when the user says "问 grok", "问 gemini", "ask grok", "ask gemini", "用 grok 查", "grok 怎么看", "让 gemini 分析", or wants a second opinion from Grok/Gemini without paying API tokens. Routes to opencli which drives the already-logged-in Chrome session.
Ask Grok or Gemini via opencli (browser-session based, no API cost). Use when the user says "问 grok", "问 gemini", "ask grok", "ask gemini", "用 grok 查", "grok 怎么看", "让 gemini 分析", or wants a second opinion from Grok/Gemini without paying API tokens. Routes to opencli which drives the already-logged-in Chrome session.
Ask via OpenCLI (Grok / Gemini)
Use opencli to query Grok or Gemini through the user's already-logged-in Chrome session. Zero API cost, fully driven by the existing browser session.
User explicitly asks to query Grok or Gemini ("问一下 grok", "用 gemini 查", "让 grok 看看")
User wants a second opinion / external advisor from Grok or Gemini
User says "ask grok"/"ask gemini" or mentions these two as the information source
When NOT to trigger:
User wants to use native Gemini CLI (gemini -p) → use ask-gemini skill instead
User wants Claude's own opinion → answer directly
User wants other AI (Doubao, Yuanbao, ChatGPT, etc.) → use other tools
Model selection
If the user doesn't specify which model, default to Grok for real-time / social / news / tech-trend questions, and for long-form reasoning, research, coding, and document analysis.
Gemini
If still ambiguous, ask the user which one to use before sending.
--timeout 90 — extend timeout for long answers (default 60s)
Parallel dispatch (for multi-AI workflows)
To call grok and gemini in parallel from Claude Code, use the Bash tool with run_in_background=true for both calls in the same assistant turn. Claude Code will launch them concurrently and wait for task-completion notifications.
For deep cross-validation workflows (N internal sub-agents + 2 external AIs, confidence-tiered synthesis), use the multi-ai-research skill instead — it covers the full orchestration pattern.
Prerequisites (check before running)
opencli installed: opencli --version should print a version.
Chrome running with Browser Bridge extension loaded — the extension is installed from ~/Desktop/code/AI/tools/opencli/extension/ via chrome://extensions → Developer mode → Load unpacked. If the extension is missing, opencli doctor will flag it.
Logged in to the target site:
Grok: grok.com must be logged in
Gemini: gemini.google.com must be logged in
Daemon connectivity: run opencli doctor once per session if unsure.
If any check fails, do not fall back to another tool silently. Report the missing prerequisite and ask the user to fix it (per U-23 no-silent-degradation rule).
Execution protocol
Decide which model to use (Grok vs Gemini). If unclear, ask.
Build the prompt. Be explicit — these browser UIs have no system prompt so the full question must be self-contained.
Run the opencli command with --format plain.
Capture the response. If the command fails with a session/auth/challenge message, stop and surface the error to the user. Do NOT retry blindly.
Save an artifact (see below).
Report back to the user: model used, summary of response, and path to the artifact.
Error handling
opencli will emit specific errors like:
"Not logged in to grok.com" → ask user to log in
"Composer not found" → DOM changed, ask user to open the site manually and retry
"Session gated / challenge" → user needs to solve the challenge in Chrome manually
Timeout → offer to retry with a longer --timeout
Do not swallow these. Report verbatim with the fix action.
Artifact requirement
After every successful call, save a markdown artifact to:
