Changelog Management Agent
Build Agents, Not Infrastructure

Changelog Management Agent

This agent receives GitHub webhooks when releases or tags are created and automatically updates changelogs across repositories using Devin AI and LLM processing.

Overview

The Changelog Management Agent acts as a smart bridge between GitHub webhooks and Devin AI. When a repository has a new release or tag created, this agent:

1. Receives the webhook from GitHub
2. Uses LLM processing to analyze the payload and determine event type, repository, and version
3. Dynamically generates a comprehensive prompt for Devin AI
4. Sends the prompt to Devin AI, which then handles the changelog generation
5. Monitors the response and status

Configuration

Create a .env file with your Devin API key and other required configuration:

DEVIN_API_KEY=your_devin_api_key_here
GITHUB_WEBHOOK_SECRET=your_github_webhook_secret_here

GitHub Webhook Setup

Configure a webhook in each repository with:

• Payload URL: Your deployed agent URL
• Content type: application/json
• Events: Releases and Tags

LLM-Driven Approach

This agent leverages AI in multiple ways:

1. Webhook Analysis: Uses an LLM to parse and extract meaning from webhook payloads
2. Prompt Generation: Dynamically creates detailed prompts for Devin AI
3. Structured Output: Uses Zod schemas to ensure reliable data extraction
4. Changelog Generation: Delegates the actual creation to Devin AI

Documentation

For comprehensive documentation on Agentuity, visit our documentation site at agentuity.dev.

Development

This agent is built using the Agentuity SDK.

# Install dependencies
bun install

# Run the agent locally
agentuity dev

Deployment

# Deploy with Agentuity
agentuity deploy

How It Works

1. GitHub webhook triggers the agent when a release/tag is created
2. The LLM analyzes the webhook payload to determine:
   • If it's a valid release/tag event
   • Which repository it relates to
   • The version number
   • If the repository is supported
3. A second LLM pass generates a detailed, repository-specific prompt
4. Devin AI processes the prompt, accessing repositories and updating changelogs
5. Both the source repository and documentation repository are updated