Agent Execution Trace
What makes a coding agent look intelligent?
Replay a recorded coding-agent trace: retrieve, edit, run, fail, repair and verify.
Step through a coding task as the agent retrieves files, follows references, edits code, runs tests, observes a failure, repairs the implementation and verifies the result. The model is only one part of the loop.
A recorded execution trace from a simulated coding agent. Nothing is executed in this browser session.
Related research
This archive contains independent research, experiments, architecture studies, and technical observations. Some findings are exploratory and should not be interpreted as peer-reviewed scientific conclusions unless explicitly stated. Architecture studies of third-party products are based on publicly observable behaviour, published technical information, and engineering analysis. They are not claims of access to private source code or infrastructure.