alexgreensh/token-optimizer v5.11.27

on GitHub

First-read skeleton transparency + markdown opt-out (#79).

• Markdown is no longer skeletonized on first read — a headings-only outline drops load-bearing prose, and the edit-rate promotion gate measures "will it be edited" rather than "will the prose be needed." Markdown first-reads now return full content (still measured in shadow). Code cohorts (Python/TypeScript signatures+imports, structure-preserving) are unchanged.
• First-read flags are now visible and toggleable on the v5 status screen and dashboard Manage tab (v5_first_read_shadow master switch, v5_first_read_active serve-vs-measure).
• Config parsing fixed — a config.json string "0"/"false"/"off" now disables consistently with the env path (was evaluating truthy via bool()); the structure-map tri-state still honors only beta; unrecognized values warn without echoing the value.
• Archive records now carry file_path so a degraded read is self-identifiable.
• Docs corrected to describe first-read substitution accurately (code-only, with expand/range/edit recovery).