A local tool checks whether Claude Code gets worse in long sessions
contextrot checks Claude Code session history to see whether long work sessions actually become less reliable. Claude Code saves each session as a JSONL log, and the tool reads those . It looks for signs such as failed edits, missed edits, repeated retries, files being read again, self-corrections, and tool errors.
It then compares those signs with how full the is to see whether failures rise over time. The result can be one of four judgments: clear , edge rot, no measurable , or not enough data. In the shared case, the tool found no measurable , and the failure rate stayed almost flat as the filled.
on the user’s computer.
Key points
- contextrot analyzes local Claude Code .
- It checks concrete failure signs like retries, missed edits, and tool errors.
- It compares those signs against how full the is.
- It can report clear rot, edge rot, no measurable rot, or not enough data.
- In this case, longer context did not measurably increase failures.