Sipcode: A context hygiene tool for Claude Code that reduces noise and improves AI response quality

What I measure directly is per-rewriter signal kept (every rewriter declares an integrity score on each fire), tool-output savings on a locked 20-task benchmark (62.6% median, range 37.4% to 80.6%, reproducible via sipcode benchmark from the repo)

The 29% quality lift number is Anthropic's published research, not mine. I am careful not to claim Sipcode users specifically see 29%.

Caps verbose tool output and dedupes same-session re-reads so the model sees signal, not noise.

Proof: 62.6% median tool-output savings on a locked 20-task benchmark. MIT.