description	allowed-tools
Full reference guide for Gemini CLI syntax and usage patterns	Read

Gemini CLI Reference

Use the Gemini CLI with its massive context window when analyzing large codebases or multiple files that might exceed context limits.

Command Syntax

gemini -m <model> -p "<prompt>"

Available Models

Model	Description
gemini-2.5-pro	Complex tasks, deep reasoning (recommended default)
gemini-2.5-flash	Balance of speed and reasoning

Including Context

Method 1: Include directories

Use --include-directories to add directories to the workspace context:

gemini -m gemini-2.5-pro --include-directories src,docs -p "Explain the architecture"

• Comma-separated list of directories
• Maximum 5 directories
• Can use multiple flags: --include-directories src --include-directories lib

Method 2: Include all files

Use --all-files (or -a) to recursively include all files in current directory:

gemini -m gemini-2.5-pro --all-files -p "Summarize this project"

Method 3: Pipe file content via stdin

Pipe specific file content directly to Gemini:

cat src/main.py | gemini -p "Explain this file"
cat README.md | gemini -p "Summarize this documentation"

Combine multiple files:

cat src/*.py | gemini -p "Review this Python code for security issues"

Configuration Options

Option	Description	Example
--prompt, -p	Run in headless mode	gemini -p "query"
--model, -m	Specify the Gemini model	gemini -m gemini-2.5-flash
--include-directories	Include additional directories	gemini --include-directories src,docs
--all-files, -a	Include all files in current dir	gemini -a -p "query"
--output-format	Output format (text, json, stream-json)	gemini -p "query" --output-format json
--debug, -d	Enable debug mode	gemini -p "query" --debug

Example Commands

Analyze a directory:

gemini -m gemini-2.5-pro --include-directories src -p "Explain the architecture of this codebase"

Review code for issues:

cat src/auth.py | gemini -p "Review this authentication code for security issues"

Full project analysis:

gemini -m gemini-2.5-pro --all-files -p "Give me an overview of this project"

Multiple directories:

gemini -m gemini-2.5-pro --include-directories src,tests,docs -p "Analyze test coverage"

Get JSON output:

gemini -p "Explain this code" --output-format json | jq '.response'

Code Review Examples

Task	Command
Security review	cat src/api/*.py | gemini -p "Check for security vulnerabilities"
Code quality	gemini --include-directories src -p "Review code quality and suggest improvements"
Architecture analysis	gemini --all-files -p "Explain the architecture and identify issues"
Test coverage	gemini --include-directories src,tests -p "Analyze test coverage gaps"
Documentation check	gemini --include-directories src -p "Are there missing docstrings or comments?"

When to Use Gemini CLI

• Analyzing entire codebases or large directories
• Comparing multiple large files
• Understanding project-wide patterns or architecture
• Current context window is insufficient
• Working with files totaling more than 100KB
• Verifying features, patterns, or security measures across codebase

Important Notes

• Gemini CLI runs from your current working directory
• Pipe file content for single-file analysis
• Use --include-directories for multi-directory analysis
• JSON output with --output-format json is useful for scripting