home / skills / bahayonghang / my-claude-code-settings / gemini

gemini skill

/content/skills/ai-llm-skills/gemini

This skill coordinates Gemini CLI to generate, review, and analyze code leveraging current web data and codebase insights for faster outcomes.

npx playbooks add skill bahayonghang/my-claude-code-settings --skill gemini

Review the files below or copy the command above to add this skill to your agents.

Files (6)
SKILL.md
6.4 KB
---
name: gemini-cli
description: Wield Google's Gemini CLI as a powerful auxiliary tool for code generation, review, analysis, and web research. Use when tasks benefit from a second AI perspective, current web information via Google Search, codebase architecture analysis, or parallel code generation. Also use when user explicitly requests Gemini operations.
category: development-tools
tags: [cli, gemini, code-generation, web-search, google]
allowed-tools:
  - Bash
  - Read
  - Write
  - Grep
  - Glob
---

# Gemini CLI Integration Skill

This skill enables Claude Code to effectively orchestrate Gemini CLI (v0.27.0+) for code generation, review, analysis, and specialized tasks. The default model is Auto routing (automatically selects between Flash and Pro based on task complexity).

## When to Use This Skill

### Ideal Use Cases

1. **Second Opinion / Cross-Validation**
   - Code review after writing code (different AI perspective)
   - Security audit with alternative analysis
   - Finding bugs Claude might have missed

2. **Google Search Grounding**
   - Questions requiring current internet information
   - Latest library versions, API changes, documentation updates
   - Current events or recent releases

3. **Codebase Architecture Analysis**
   - Use Gemini's `codebase_investigator` tool
   - Understanding unfamiliar codebases
   - Mapping cross-file dependencies

4. **Parallel Processing**
   - Offload tasks while continuing other work
   - Run multiple code generations simultaneously
   - Background documentation generation

5. **Specialized Generation**
   - Test suite generation
   - JSDoc/documentation generation
   - Code translation between languages

### When NOT to Use

- Simple, quick tasks (overhead not worth it)
- Tasks requiring immediate response (rate limits cause delays)
- When context is already loaded and understood
- Interactive refinement requiring conversation

## Core Instructions

### 1. Verify Installation

```bash
command -v gemini || which gemini
```

### 2. Basic Command Pattern

```bash
gemini "[prompt]" --approval-mode yolo -o text 2>&1
```

Key flags:
- `--approval-mode yolo`: Auto-approve all tool calls (replaces deprecated `--yolo`)
- `-o text`: Human-readable output
- `-o json`: Structured output with stats
- `-m gemini-2.5-flash`: Use faster model for simple tasks
- `-m auto`: Smart routing (default) — auto-selects Flash or Pro based on complexity

### 3. Critical Behavioral Notes

**Approval Mode Behavior**: `--approval-mode yolo` auto-approves tool calls but does NOT prevent planning prompts. Gemini may still present plans and ask "Does this plan look good?" Use forceful language:
- "Apply now"
- "Start immediately"
- "Do this without asking for confirmation"

> **Note:** `--yolo` / `-y` is deprecated. Use `--approval-mode yolo` instead. Other modes: `default`, `auto_edit`.

**Rate Limits**: Free tier limits apply. CLI auto-retries with exponential backoff. Expect messages like "quota will reset after Xs".

### 4. Output Processing

For JSON output (`-o json`), parse:
```json
{
  "response": "actual content",
  "stats": {
    "models": { "tokens": {...} },
    "tools": { "byName": {...} }
  }
}
```

## Quick Reference Commands

### Code Generation
```bash
gemini "Create [description] with [features]. Output complete file content." --approval-mode yolo -o text
```

### Code Review
```bash
gemini "Review [file] for: 1) features, 2) bugs/security issues, 3) improvements" -o text
```

### Bug Fixing
```bash
gemini "Fix these bugs in [file]: [list]. Apply fixes now." --approval-mode yolo -o text
```

### Test Generation
```bash
gemini "Generate [Jest/pytest] tests for [file]. Focus on [areas]." --approval-mode yolo -o text
```

### Documentation
```bash
gemini "Generate JSDoc for all functions in [file]. Output as markdown." --approval-mode yolo -o text
```

### Architecture Analysis
```bash
gemini "Use codebase_investigator to analyze this project" -o text
```

### Web Research
```bash
gemini "What are the latest [topic]? Use Google Search." -o text
```

### Faster Model (Simple Tasks)
```bash
gemini "[prompt]" -m gemini-2.5-flash -o text
```

## Error Handling

### Rate Limit Exceeded
- CLI auto-retries with exponential backoff
- Use `-m gemini-2.5-flash` for lower priority tasks (different quota)
- Use Auto routing (default) to let the system optimize model selection
- Run in background for long operations

### Command Failures
- Check JSON output for detailed error stats
- Verify Gemini is authenticated: `gemini --version`
- Check `~/.gemini/settings.json` for config issues

### Validation After Generation
Always verify Gemini's output:
- Check for security vulnerabilities (XSS, injection)
- Test functionality matches requirements
- Review code style consistency
- Verify dependencies are appropriate

## Integration Workflow

### Standard Generate-Review-Fix Cycle

```bash
# 1. Generate
gemini "Create [code]" --approval-mode yolo -o text

# 2. Review (Gemini reviews its own work)
gemini "Review [file] for bugs and security issues" -o text

# 3. Fix identified issues
gemini "Fix [issues] in [file]. Apply now." --approval-mode yolo -o text
```

### Background Execution

For long tasks, run in background and monitor:
```bash
gemini "[long task]" --approval-mode yolo -o text 2>&1 &
# Monitor with BashOutput tool
```

## Gemini's Unique Capabilities

These tools are available only through Gemini:

1. **google_web_search** - Real-time internet search via Google
2. **codebase_investigator** - Deep architectural analysis (experimental agent, may require opt-in)
3. **save_memory** - Cross-session persistent memory

## Model Selection

| Model | Use Case |
|-------|----------|
| `auto` (default) | Smart routing — auto-selects Flash or Pro based on task complexity |
| `gemini-2.5-pro` | Complex reasoning tasks |
| `gemini-2.5-flash` | Quick tasks, lower latency |
| `gemini-3-pro-preview` | Latest capabilities (requires Google AI Ultra or paid API key) |
| `gemini-3-flash-preview` | Latest Flash model (requires Preview Features enabled) |

## Configuration

### Project Context (Optional)

Create `.gemini/GEMINI.md` in project root for persistent context that Gemini will automatically read.

### Session Management

List sessions: `gemini --list-sessions`
Resume session: `echo "follow-up" | gemini -r [index] -o text`

## See Also

- `reference.md` - Complete command and flag reference
- `templates.md` - Prompt templates for common operations
- `patterns.md` - Advanced integration patterns
- `tools.md` - Gemini's built-in tools documentation

Overview

This skill integrates the Gemini CLI as an auxiliary tool for code generation, review, analysis, and web-grounded research. It lets you run Gemini from the command line to get a second AI perspective, access current web information via Google Search, and perform deep codebase investigations. Use it when tasks benefit from parallel processing, model diversity, or Gemini-specific tools.

How this skill works

The skill issues gemini CLI commands with recommended flags and parses text or JSON output. It can run generation, review, bug-fix, test and documentation workflows, and invoke Gemini tools like google_web_search and codebase_investigator. For automation it uses --approval-mode yolo to auto-approve tool calls and -o json for structured results to inspect stats and tooling usage.

When to use it

  • Ask for a second AI opinion or cross-validation of code and designs.
  • Need current internet grounding (library versions, docs, recent releases).
  • Analyze unfamiliar codebases or map cross-file dependencies.
  • Generate tests, docs, or translate code in parallel with other work.
  • Offload longer or background tasks that can run outside the main session.

Best practices

  • Use --approval-mode yolo and forceful language (e.g., "Apply now") for automated tool actions.
  • Prefer -o json for programmatic parsing and -o text for human-readable output.
  • Use model flags (-m auto / gemini-2.5-flash / gemini-2.5-pro) based on task complexity.
  • Validate all generated code: run tests, review for security issues, and check dependencies.
  • Run long or rate-limited jobs in background and monitor results via logs or session resume.

Example use cases

  • Generate a full file implementation and then run Gemini to review and patch issues in a generate-review-fix cycle.
  • Use google_web_search to gather the latest API changes and update code accordingly.
  • Run codebase_investigator to get a dependency map and suggest refactor targets for a legacy project.
  • Produce comprehensive Jest or pytest suites for a module and apply fixes automatically.
  • Translate a module from one language to another and validate type and API compatibility.

FAQ

How do I avoid Gemini asking for plan approval?

Include --approval-mode yolo and explicit instructions like "Start immediately" or "Apply now" to reduce interactive confirmations.

When should I prefer gemini-2.5-flash?

Use gemini-2.5-flash for simple, lower-latency tasks or when you want to conserve higher-tier quota.