← AI GitHub

AI GitHub

Aider for terminal-based pair programming

Make model-assisted edits in a Git checkout and review the result as a diff.

Coding AssistantsPublished 2 min read

Key Takeaways

  • Aider uses repository context and explicitly selected files to help a model propose edits.
  • A local Git checkout, a configured model and a known validation command make a useful starting point.
  • Use a clean branch with a small bug and a known test.
  • An agent can modify files beyond the line mentioned in the prompt.

How it works

Aider uses repository context and explicitly selected files to help a model propose edits. Its Git integration keeps changes visible and gives you a place to compare successive attempts.

What you need

A local Git checkout, a configured model and a known validation command make a useful starting point. Keep an untouched reference branch for inspecting changes.

What to try first

Use a clean branch with a small bug and a known test. Add the files that define the behavior, ask for one fix, then review the diff and run the test yourself.

Include a regression test that fails on the original branch. That shows whether the agent repaired behavior or simply found a way to keep an existing test green.

Where to look for failures

An agent can modify files beyond the line mentioned in the prompt. Keep secrets out of model context, check commit behavior and distinguish your pre-existing edits from the generated ones.

Related Franklin pages

Use the same clean checkout and failing test to compare terminal agents. Look for correct file selection, a minimal diff and evidence from the test run.

For a library of development workflows that can complement a coding agent, see Franklin’s Superpowers profile. Superpowers

Project documentation

Try it for your workflow

Explore the project’s documentation for setup and usage, or save this page to a collection for later.

Read the project documentation ↗