contextrot measures Claude Code degradation using session logs
The contextrot utility analyzes the exact context fill level at which an agent begins to fail. It reads local JSONL transcripts that Claude Code already writes to disk, requiring no API keys or network requests.
The tool extracts five failure signals at each step: failed code edits, repeated calls after errors, re-reading already processed files, self-corrections, and tool errors. It then correlates these with context fill levels, calculating statistics using 95% Wilson score intervals.
The report provides one of four verdicts: degradation detected, edge degradation at the window limit, no degradation, or insufficient data. You can run it via uvx contextrot or pip3 install contextrot. Support for Codex CLI, OpenCode, and Gemini CLI is planned. The project is under the MIT license.
Up next
Over 100 authors sue Anthropic for $75 million
Over 100 authors, including Atari co-founder Nolan Bushnell, are suing Anthropic for $75 million. They allege the company used 500 pirated books to train Claude and are seeking $150,000 per title. The plaintiffs opted out of a prior settlement that paid roughly $3,000 per work. They now claim Anthropic is liable for both downloading and distributing their books.
Related stories
- Intent-lens spells out Claude Code calls before they run
- Codedict cuts Claude Code tokens 63%, but cost only 26%
- claude-files-pane shows Claude Code’s file reads live
- lcc cut one Claude Code session 49.5% without a summary
- catnap keeps Claude Code from compacting mid-task
- Script checks whether your network breaks Claude API uploads
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.