gemini

Launch Gemini CLI sessions for code reviews and consultations with approval controls.

1|Updated Mar 11, 2026

Search Tags:#claude #multi-agent #gemini #session-management #code-review #structured-output

Ready to run

npx skills add https://github.com/andreahaku/gemini_cli_skill --skill gemini-andreahaku

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: gemini
Source: https://github.com/andreahaku/gemini_cli_skill/tree/main/skill/gemini
Command: npx skills add https://github.com/andreahaku/gemini_cli_skill --skill gemini-andreahaku

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires gemini, python3, git, file, sed, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill centralizes Gemini CLI usage so Claude Code can consult, review, and debate across large contexts without manual session juggling or unsafe commands.

Core Features & Use Cases

Session-aware consultations: Run ask scripts that resume, list, or start new Gemini sessions with approval controls and depth flags.
Code reviews and debates: Launch review wrappers, structured JSON outputs, and debate scripts to surface critiques and cross-model synthesis for architecture or security audits.
Multi-agent orchestration: Track threads, link sessions, and coordinate worker mode outputs so Gemini contributes alongside Codex or other Claude agents in large-scale workflows.

Quick Start

Ask Claude to use the gemini skill and describe the code review or consultation you want Gemini to perform.

Frequently Asked Questions about gemini

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I run a Gemini CLI code review on a large git-tracked repository?▼

You can run a Gemini CLI code review on a large git-tracked repository by using the skill's review wrappers, which provide structured JSON outputs and session persistence for deep architecture checks and security audits.

What is multi-agent session management for cross-model code debates?▼

Multi-agent session management for cross-model code debates coordinates Gemini alongside other agents using debate scripts to surface critiques and synthesize architecture or security findings without manual session juggling.

How do I handle Gemini CLI rate limits during large-context consultations?▼

Handle Gemini CLI rate limits during large-context consultations through the skill's built-in fallback retry logic, which automatically manages API constraints when deep reasoning tasks hit request limits.

Does this approach require Python3 and git for multi-agent orchestration?▼

Yes, this approach requires Python3 and git for multi-agent orchestration, as the wrappers depend on these tools to manage repository-level contexts, track threads, and coordinate worker mode outputs.

Can I use Gemini for structured JSON outputs in multi-agent workflows?▼

You can use Gemini for structured JSON outputs in multi-agent workflows by passing model and depth flags through the skill's consultation scripts, ensuring responses meet precise formatting requirements for downstream agent processing.

What are the limitations of using CLI wrappers for repository-level security audits?▼

Limitations of using CLI wrappers for repository-level security audits include dependency on external rate limits that trigger fallback retries, and the need for safe approval modes to prevent unsafe commands during automated consultations.