gemini-chat-and-reasonlisted

# Gemini Chat and Reason Use this skill when you want Gemini to serve as a reasoning partner: reviewing code, critiquing a design, stress-testing a decision, or providing a second opinion on anything Claude has produced or is about to produce. ## When to use this skill - **Code review:** You have written a non-trivial function or module and want a second set of eyes before committing. - **Design critique:** You are proposing an architecture, API contract, or data model and want pushback. - **Sanity check:** You are about to take an irreversible action (deploy, delete, migrate) and want confirmation the plan is sound. - **Perspective shift:** Claude's reasoning has gone deep down one path; Gemini can surface blind spots. - **Multi-turn dialogue:** You need to iterate on a draft or decision through several back-and-forth exchanges. ## MCP tools | Tool | Purpose | |---|---| | `mcp__gemini__gemini_generate` | Single-shot text generation, best for one-off questions or reviews | | `mcp__gemini__gemini_chat` | Multi-turn conversation, best for iterative critique or dialogue | ## Model selection guidance - For routine second opinions and code reviews, use the default model (Gemini 2.5 Flash or equivalent); it is fast and cheap. - For complex architectural decisions or lengthy documents requiring deep reasoning, request `gemini-2.5-pro` explicitly via the `model` parameter. - Avoid `gemini-2.5-pro` on short, low-stakes queries; the cost uplift is not justified. ## Usage patter