introducing proof! it's an agent-native markdown editor with provenance tracking. keeps track of what was written by you vs. AI, and lets you mark which parts of a plan document (or any .md file) you've reviewed, and how closely you've reviewed it. red = written by AI, no review. yellow = i've skimmed it green = i endorse this / i wrote this i think it'll be great as the default plan editor in Claude Code and Codex. releasing internally @every this week and if it goes well, will release more broadly