Paul-Oriented Programming with Aider

A specialized Aider configuration for developing the Paul-Oriented Programming Language (PaulTalk) using Google's Gemini 2.5 models with unified diff format and aggressive file watching.

What This Configuration Does

This skill configures Aider to work optimally with the PaulTalk programming language project, leveraging Gemini 2.5's high-context capabilities and unified diff format for precise code modifications.

Configuration Details

**Models:**

**Edit Format:**

**Context & Performance:**

**Monitoring:**

**Language & Guidelines:**

How to Use

1. **Install Aider:**

```bash

pip install aider-chat

```

2. **Save Configuration:**

Create `.aider.conf.yml` in your project root with this content:

```yaml

model: gemini/gemini-2.5-pro

editor-model: gemini/gemini-2.5-pro

weak-model: gemini/gemini-2.5-flash

edit-format: udiff

editor-edit-format: udiff

max-chat-history-tokens: 1000000

map-tokens: 3000

chat-language: English

cache-prompts: true

map-refresh: always

watch-files: true

verbose: true

read:

- AiderGuidelines.md

```

3. **Set API Key:**

```bash

export GEMINI_API_KEY=your_api_key_here

```

4. **Launch Aider:**

```bash

aider

```

The configuration will be automatically loaded from `.aider.conf.yml`.

5. **Start Coding:**

```

> Add a new parser function for handling string literals

> Refactor the tokenizer to support multi-line comments

> Fix the type inference bug in src/typechecker.ts

```

When to Use This Skill

Key Features

**High Context Window:**

The 1M token limit allows Aider to maintain extensive conversation history, ideal for complex programming language development discussions.

**Unified Diff Format:**

The `udiff` format provides clear before/after views of changes, reducing errors and improving code review.

**Automatic File Watching:**

Changes made outside Aider are detected immediately, keeping the AI's understanding of the codebase current.

**Project Guidelines Integration:**

The `read: [AiderGuidelines.md]` directive ensures the AI follows project-specific conventions automatically.

Configuration Options Explained

| Option | Value | Purpose |

|--------|-------|---------|

| `model` | gemini/gemini-2.5-pro | Primary AI model for code generation |

| `editor-model` | gemini/gemini-2.5-pro | Model for editing existing code |

| `weak-model` | gemini/gemini-2.5-flash | Faster model for simple operations |

| `edit-format` | udiff | Unified diff format for changes |

| `max-chat-history-tokens` | 1000000 | Maximum conversation context |

| `map-tokens` | 3000 | Repository structure map size |

| `cache-prompts` | true | Cache prompts for performance |

| `map-refresh` | always | Keep repo map up-to-date |

| `watch-files` | true | Monitor external file changes |

| `verbose` | true | Detailed logging output |

Notes

Related Resources