Inspiration
We ran into problems with collaborating and coordinating larger projects in github, especially when handling pull requests from different branches and dealing with merge errors. We thought that this was the perfect chance to implement a multiagent model that acts as a github assistant that is specifically designed to deal with these problems.
What it does
Automates commits, creates branches and allocates workloads. It will be a vscode extension.
How we built it
Multiagent framework using langgraph in python.
Challenges we ran into
Turning our project into a vscode extension. Orchestrating our different agents.
Accomplishments that we're proud of
Implemented multiple features:
- Task Creator
- Dependencies Mapper
- Branch Creator
- Task Allocation
- Completion checker
- Git committer
- Slack Agent We successfully coded an orchestrator that works with the features we implemented.
What we learned
How to build a multiagent system using langgraph and langchain. How to use CAIPE (although we did not implement it).
What's next for Pacer
Jira integration. Agent feedback.
Log in or sign up for Devpost to join the conversation.