Aider Configuration

Configure Aider AI pair programming assistant with customized models, git workflows, editor preferences, and interaction settings for your development environment.

What This Skill Does

This skill helps you set up a comprehensive Aider configuration file (`.aider.conf.yml`) that controls how Aider behaves in your project. It enables you to customize AI models, automate git commits, configure output styling, set up linting and testing workflows, and personalize the coding assistant experience.

Instructions

When the user requests Aider configuration setup:

1. **Determine Configuration Location**

- Ask where to place the config: project root (`.aider.conf.yml`) or home directory (`~/.aider.conf.yml`)

- Project-level configs override home directory configs

2. **Gather Model Preferences**

- Ask which model to use for main chat (e.g., `gemini/gemini-2.5-pro`, `claude-3-7-sonnet-20250219`, `gpt-4o`)

- Ask about weak model for summaries (optional, defaults based on main model)

- Ask about editor model for quick edits (optional, faster models recommended)

- Set `edit-format` if user has preference (diff, whole, or architect mode)

3. **Configure Git Integration**

- Ask if auto-commits should be enabled (`auto-commits: true/false`)

- Determine attribution preference (`attribute-author`, `attribute-committer`, or `attribute-co-authored-by`)

- Ask if commit messages should be prefixed with 'aider:' (`attribute-commit-message-author` or `attribute-commit-message-committer`)

- Configure dirty commits handling (`dirty-commits: true/false`)

4. **Set Up Linting and Testing**

- Ask if auto-linting should be enabled (`auto-lint: true/false`)

- If yes, gather lint commands for relevant languages (e.g., `python: flake8 --select=...`)

- Ask about test command if applicable (`test-cmd`)

- Determine if auto-testing is desired (`auto-test: true/false`)

5. **Configure Output and Interface**

- Ask about terminal theme (`dark-mode` or `light-mode`)

- Determine if vim mode is preferred (`vim: true/false`)

- Set code theme for markdown output if desired (`code-theme`)

- Configure notification preferences (`notifications: true/false`)

6. **Set Up Read-Only Files**

- Ask if there are configuration files that should be read-only (e.g., `tsconfig.json`, `package.json`, `.prettierrc`)

- Add patterns for files that provide context but shouldn't be edited

7. **Configure Advanced Settings**

- Repo map tokens (`map-tokens`) for codebase understanding

- Cache settings (`cache-prompts`, `cache-keepalive-pings`)

- Analytics preferences (`analytics-disable: true` recommended)

- Editor preference for `/editor` command

8. **Create Configuration File**

- Write `.aider.conf.yml` with selected settings

- Include comments explaining key options

- Omit unneeded settings (don't include every possible option)

9. **Add .aiderignore if Needed**

- Ask if certain files/directories should be ignored

- Create `.aiderignore` file with patterns (similar to `.gitignore`)

10. **Verify and Document**

- Confirm configuration matches user's workflow

- Explain how to modify settings later

- Note that API keys should go in `.env` file, not YAML config

Configuration Examples

Minimal Configuration

```yaml

model: claude-3-7-sonnet-20250219

dark-mode: true

auto-commits: false

analytics-disable: true

```

Full-Featured Configuration

```yaml

model: gemini/gemini-2.5-pro

weak-model: gemini/gemini-2.5-flash

editor-model: gemini/gemini-2.0-flash

dark-mode: true

vim: true

auto-commits: true

auto-lint: true

lint-cmd:

- "python: ruff check --fix"

- "javascript: eslint --fix"

auto-test: false

test-cmd: "npm test"

attribute-co-authored-by: true

analytics-disable: true

editor: nvim

read:

- tsconfig.json

- package.json

- .prettierrc

```

Important Notes

Common Use Cases

1. **Team Standardization**: Create project-level config for consistent Aider behavior across team

2. **Personal Preferences**: Set up home directory config with your preferred editor, colors, and vim mode

3. **CI/CD Integration**: Disable auto-commits and prompts for automated environments

4. **Cost Optimization**: Use cheaper models for weak/editor models while keeping powerful model for main chat

5. **Compliance**: Configure attribution settings to meet your organization's requirements