Context Engineering Template for Gemini CLI

A comprehensive template to jumpstart Context Engineering with Gemini CLI—ensuring your AI coding assistant has all the context needed for end-to-end feature implementation.

Context Engineering with Gemini CLI is 10x smarter than prompt engineering and 100x more reliable than vibe coding.

🚀 Quick Start

# 1. Clone this template
git clone https://github.com/coleam00/context-engineering-intro-GeminiCLI.git
cd context-engineering-intro

# 2. Customize project guidelines (optional)
# Edit GEMINI.md to add your project-specific conventions

# 3. Add illustrative examples (highly recommended)
# Place relevant code samples in the examples/ folder

# 4. Define your initial feature request
# Edit INITIAL.md with detailed requirements

# 5. Generate a comprehensive PRP (Product Requirements Prompt)
# Using Gemini CLI, run:

gemini context generate-prp INITIAL.md

# 6. Execute the PRP to implement your feature
# Using Gemini CLI, run:

gemini context execute-prp PRPs/your-feature-name.md

📚 Table of Contents

• What is Context Engineering?
• Template Structure
• Step-by-Step Guide
• Writing Effective INITIAL.md Files
• The PRP Workflow
• Using Examples Effectively
• Best Practices

What is Context Engineering?

Context Engineering shifts the focus from single-shot prompts to a structured system of rules, examples, and validations:

Prompt Engineering vs Context Engineering

Prompt Engineering:

• Crafting clever wording
• Limited to phrasing
• Like leaving a sticky note

Context Engineering with Gemini CLI:

• Full framework: docs, examples, rules, tests
• End-to-end feature pipelines
• Like directing a full movie production

Why It Matters

1. Fewer AI Failures: Most issues stem from missing context, not model errors.
2. Consistent Results: Enforce your code style and patterns.
3. Supports Complex Flows: Multi-step features handled seamlessly.
4. Automated Validation: Built‑in tests and linting loops ensure quality.

Template Structure

context-engineering-intro/
├── .gemini/
│   ├── commands/
│   │   ├── generate-prp.md    # PRP generation logic
│   │   └── execute-prp.md     # PRP execution logic
│   └── settings.local.json    # Gemini CLI permissions
├── PRPs/
│   ├── templates/
│   │   └── prp_base.md        # Base PRP template
│   └── EXAMPLE_multi_agent_prp.md  # Sample complete PRP
├── examples/                   # Your code examples (critical!)
├── GEMINI.md                   # Global project rules for Gemini CLI
├── INITIAL.md                  # Template for new feature requests
├── INITIAL_EXAMPLE.md          # Example feature request
├── INITIAL_COMPLEX.md          # More comprehensive document for more complex projects
└── README.md                   # This file (updated for Gemini CLI)

Step-by-Step Guide

1. Define Global Rules (GEMINI.md)

GEMINI.md holds project‑wide conventions for your AI assistant:

• Awareness: Task and planning docs
• Structure: Module splits, file size limits
• Testing: Unit test templates, coverage thresholds
• Style: Language and formatting rules
• Docs: Docstring and commenting practices

Customize or use the template as‑is for your project needs.

2. Draft Your INITIAL.md

Describe your new feature:

## FEATURE:
[Clear, detailed description]

## EXAMPLES:
[List example files and patterns]

## DOCUMENTATION:
[Links to API docs, schemas, guides]

## OTHER CONSIDERATIONS:
[Auth, rate limits, gotchas]

See INITIAL_EXAMPLE.md for guidance.

Use INITIAL_COMPLEX.md for more complex feature

## High-Level Objective
[What are you building?]

## Mid-Level Objectives
[What are the steps to get there?]

## Implementation Notes
[Tech notes, dependencies, file structure, etc.]

## Context
- Beginning Context: [Which files exist?]
- Ending Context: [What files will exist after?]

## Low-Level Tasks
1. [Detailed implementation steps]

## EXAMPLES:
[List example files and patterns]

## DOCUMENTATION:
[Links to API docs, schemas, guides]

## OTHER CONSIDERATIONS:
[Auth, rate limits, gotchas]

Before running the CLI, rename INITIAL_COMPLEX.md to INITIAL.md so Gemini can use it.

3. Generate the PRP

PRPs (Product Requirements Prompts) combine context, implementation steps, validation gates, and tests.

Run:

gemini context generate-prp INITIAL.md

The CLI will:

1. Parse your request
2. Analyze code patterns
3. Gather docs
4. Create PRPs/your-feature-name.md

4. Execute the PRP

After reviewing the generated PRP, implement it:

gemini context execute-prp PRPs/your-feature-name.md

The assistant will:

1. Load the PRP context
2. Plan steps
3. Implement code with validation
4. Run tests and linting
5. Iterate until success

Writing Effective INITIAL.md Files

• Be Explicit: Include all requirements.
• Reference Examples: Show what to emulate.
• Link Docs: Provide URLs and resources.
• Note Gotchas: Auth, quotas, edge cases.

The PRP Workflow

generate-prp

1. Research codebase patterns
2. Fetch docs and quirks
3. Draft implementation blueprint
4. Embed validation gates

execute-prp

1. Load PRP context
2. Create task list
3. Write and test code
4. Iterate on failures
5. Finalize feature

Using Examples Effectively

Examples in examples/ are critical:

• Structure Patterns: Modules, classes, functions
• Testing: Test file layouts, mocks, assertions
• Integration: API clients, DB connections
• CLI: Arg parsing, outputs, errors

Best Practices

1. INIT.md Clarity: No assumptions—spell it out.
2. Rich Examples: More = better.
3. Validation Gates: Tests in PRPs.
4. Leverage Docs: Include all relevant links.
5. Customize GEMINI.md: Enforce standards.

Resources

• Gemini CLI Docs
• Context Engineering Guide