Agentic AI Workflow Configuration

This skill configures Aider for autonomous AI agent workflows with intelligent defaults for local development, git integration, and context-aware code generation.

Overview

Configures Aider to work optimally with the AGENTIC_AI_WORKFLOW pattern, supporting both local Ollama models (recommended for privacy and zero cost) and cloud APIs. The configuration includes automated git commits, comprehensive context loading, and hardware-optimized model selection.

Configuration Strategy

Local Model Selection (Recommended)

The configuration provides three hardware tiers:

**Budget/Laptop (8GB VRAM)**

**Mid-Range (16GB VRAM)**

**High-End (24GB+ VRAM)**

Git Automation

Automated commit workflow with structured messaging:

Context Management

Pre-loaded documentation for informed decision-making:

Repository mapping with 2048 token context window for comprehensive understanding.

Code Quality

Implementation

Create `.aider.conf.yml` in your project root:

```yaml

AI Model Configuration

model: ollama/qwen2.5-coder:7b-instruct

editor-model: ollama/qwen2.5-coder:7b-instruct

api-base: http://127.0.0.1:11434

Git Integration

auto-commits: true

dirty-commits: true

commit-prompt: |

Generate a clear, descriptive commit message following AGENTIC_AI_WORKFLOW conventions.

Include AI_COMMIT metadata tags when appropriate.

Format: AI_COMMIT: [brief description]

- [detailed changes]

- Ref: [documentation reference if applicable]

Context and Files

read:

- docs/AGENTIC_AI_WORKFLOW.md

- DOCUMENTATION_GUIDE.md

- BOOTSTRAP_AUTONOMOUS_AGENT.md

map-tokens: 2048

map-refresh: auto

Code Quality

lint: true

test: false

auto-test: false

Output Configuration

pretty: true

show-diffs: true

show-repo-map: false

chat-history-file: .aider.chat.history.md

Additional Settings

dark-mode: auto

voice-language: en

edit-format: architect

env-file: .aider.env

stream: true

message-file: .aider.message.md

restore-chat-history: false

```

Usage

1. **Install Aider and Ollama** (for local models):

```bash

pip install aider-chat

# Install Ollama from https://ollama.ai

```

2. **Pull recommended models**:

```bash

ollama pull qwen2.5-coder:7b-instruct

```

3. **Place configuration** in project root as `.aider.conf.yml`

4. **Launch Aider**:

```bash

aider

```

Aider will automatically load the configuration, pre-read workflow documentation, and use the specified model tier.

Customization

**Switch to cloud APIs**: Uncomment cloud model lines and comment out Ollama configuration:

```yaml

model: gpt-4o

editor-model: gpt-4o-mini

api-base: http://127.0.0.1:11434 # Comment this out

```

**Adjust context files**: Modify the `read:` section to include project-specific documentation.

**Change hardware tier**: Uncomment the model pair matching your VRAM capacity.

**Enable chat history persistence**: Set `restore-chat-history: true` to continue previous sessions.

Benefits

Requirements