AI Agent Usage

Use gitinspect through MCP-compatible AI coding agents.

Prerequisites

1. Install gitinspect (see Installation)
2. Configure your AI agent (see Agent Config)

Example: Understanding a New Codebase

When you open a project in your AI agent, you can ask:

"Inspect this repository and give me an overview of the codebase structure"

The agent will call list_repo_files to see what's in the repo, then inspect_repo to read the key files.

Example: Focused Code Review

"Inspect only the Go files in this project, excluding tests and vendor"

The agent will call:

inspect_repo(
  repo=".",
  include="**/*.go",
  exclude="**/vendor/*,**/*_test.go",
  max_tokens=8000,
  strip=true
)

Example: Remote Repository Analysis

"Inspect the repository at https://github.com/go-git/go-git and summarize its structure"

The agent will call:

list_repo_files(repo="https://github.com/go-git/go-git")

Then based on the file list:

inspect_repo(
  repo="https://github.com/go-git/go-git",
  max_tokens=6000,
  strip=true
)

Example: Dependency Audit

"What dependencies does this project use?"

The agent will call inspect_repo and examine the dependencies field in the output.

Two-Step Pattern

The recommended pattern for AI agents:

1. list_repo_files — Get an overview with file sizes and priority scores
2. inspect_repo — Read specific files with a targeted token budget

This avoids wasting tokens on files the agent doesn't need.