AI USE CASES YOU CAN REUSE.English edition · Individuals, independents, small businesses
GUIDE · 6 CARDS

Claude Code and coding agent alternatives: real usage, setup and comparison

Claude Code is Anthropic’s command-line (CLI) and editor coding agent, capable of reading an entire repository, running tests, editing multiple files and calling MCP servers.

This guide gathers verified production lessons from engineering teams (Doctolib, Alan, Capgemini, GitHub) and compares leading alternatives (OpenAI Codex, Cursor, self-hosted setups).

See also the MCP server guide and Create an AI Agent Guide and l’IA agentique en production and all use cases.

Claude Code and Agent Skills (SKILL.md)

Claude Code relies on Agent Skills (SKILL.md files) to load domain instructions, scripts and review rules on demand without saturating the context window. Browse our AI agent skills directory and the official Claude API skill, each summarised with installation commands.

Claude Code, OpenAI Codex or Cursor: which one to choose?

Choose Claude Code for terminal-first multi-file refactoring with MCP servers, OpenAI Codex to delegate asynchronous tasks in a cloud sandbox, and Cursor for step-by-step interactive editing in the IDE. Regulated organisations often deploy self-hosted coding assistants.

IN THE DIRECTORY

IN THE DIRECTORY

Coding agents in production: code review, tests and GitHub

Alternatives: OpenAI Codex and multi-agent orchestration

  • Check scientific simulation code for errors with CodexSpotted · L’Etudiant — Chloë, PhD student, Université Grenoble-Alpes · Codex

    Chloë, a PhD student in applied mathematics in Grenoble, uses Codex to check her simulation code. "Whenever I run into an error I don't understand or want to make my code run faster, I use AI."

SKILLS

SKILLS

The whole library
AnthropicAI and machine learning

Reference of the API Claude and SDK Anthropic

claude-api

To be used as an up-to-date reference for building an application that calls the API Claude or an SDK Anthropic: model identifiers, parameters, streaming, tool call, MCP, agents, cache, token count and migration between models.

Demonstration
Tools: Claude Code, Claude.ai, API ClaudeLevel: Intermediateupdated 09/10/2026
SentryObservability and monitoring

Recommend permissions Claude Code

claude-settings-audit

The skill analyses a repository to offer Claude Code (settings.json) permissions for read-only commands.

Tools: Claude Code, Cursor, Cline +1Level: Intermediateupdated 05/11/2026
AnthropicAgents and skills

Build a quality MCP server

mcp-builder

Guides the creation of MCP servers that allow a model to act on external services with well-designed tools, in Python with FastMCP, or in Node and TypeScript with SDK MCP.

Demonstration
Tools: Claude Code, Claude.ai, API ClaudeLevel: Advancedupdated 28/10/2026
SentryObservability and monitoring

Make a code review according to Sentry

code-review

The skill rereads a change according to Sentry engineering practices: safety, performance, tests and design, typically on the occasion of a pull request.

Tools: Claude Code, Cursor, Cline +1Level: Intermediateupdated 19/04/2026
FREQUENTLY ASKED QUESTIONS

In brief, with the sources.

How do you install Claude Code and get started quickly?

Claude Code installs via the CLI or runs through Anthropic’s desktop app and IDE extensions. Once signed in with a Claude account or Anthropic API key, you can connect MCP servers with claude mcp add and load Agent Skills in the SKILL.md format.

What are the best alternatives to Claude Code?

Depending on your workflow, the main alternatives are OpenAI Codex (asynchronous agent in terminal and cloud, see the Codex case), Cursor and VS Code + GitHub Copilot / Agents. For regulated environments, teams like Capgemini deploy self-hosted coding assistants.