Welcome, builder! 👋
Trace Commons is running an open-theme track at this hackathon, judged on how well you worked with your AI coding agent. To enter, share the traces from your project. Your agent handles the submission, and taking part is entirely optional.
Trace Commons is an open-source, user-controlled register of AI-agent work.
Its contributor tool can discover supported agent sessions on your device, scrub them locally, and submit the redacted trace for review.
You can review the open-source code here: Trace Commons on GitHub
What is a trace?
A trace is the record your AI coding agent keeps of a session: your prompts, the agent’s responses, the tool calls and the commands it ran while you were building. Agents such as Claude Code and Codex store these locally on your device.
The track asks a simple question:
<aside> ✍️
How effectively can you use an AI coding agent while building your hackathon project?
</aside>
It is an open-theme track, so build whatever you like with Claude Code, Codex, or another supported agent. To enter, submit your project and share the scrubbed traces from building it. Your agent does both for you.
The focus is not just on the final product. Judges will also look at how effectively you worked with your agent to plan, debug, iterate, and ship.
Participating in the Trace Commons track involves sharing a scrubbed record of your AI-agent session.
What happens to your data