Changes touching this path

• a hygiene check runs at the moment that makes it true: the two silent-damage checks move into every land where --skip-tests cannot reach them, the CI net they were said to have turns out not to exist, and the ADR number is re-asked under the lock that makes the answer hold (#1082, #1105) cd3021b8 · dbf3dbe6…
• the land reads its position once and hands that one corpus to three gates that still run in ADR 0065's order, undoing the two-and-a-half-fold read the previous change caused by leaving a walk inside each gate while claiming there was only one ec75201e · dbf3dbe6…diff
• the line-ending land refusal now walks site/ and covers ts, tsx, mjs, css, html and py, the site extensions it did not cover, while skipping the site build output that site/.gitignore ignores (dist, .output, .tanstack, .nitro and site/src/wasm), so a flipped site source file no longer lands unrefused. js stays out because the site tracks none, txt stays out because the tracked run transcripts in docs/evidence/runs hold CRLF, and ps1 stays exempt. a planted CRLF file for each new extension under site/ went red on the old walk (10 passed, 3 failed), and dropping site from the roots or tsx from the extensions went red again (10 passed, 3 failed each). unskipping .output or site/src/wasm, or covering ps1, went red on the build-output and powershell pin (9 passed, 4 failed each), and covering txt turned the real-tree test red (1 passed, 1 failed). on this repo the walk now reads 1228 files where it read 1007, median 46 ms against 38 ms, and a built site passes it. AGENTS.md and the hygiene.rs module doc now say which files the refusal reads by pointing at the constants, AGENTS.md names the transcripts as an exception, and the NUL sentinel in site/test/prerender-list.test.ts is written as an escape so git reads the file as text. the site gate and the workspace suite are green (#1792) ab6c3116 · dbf3dbe6…diff

Renames are not followed. loot's tree maps a path to an address, so a rename is a delete and an add. This list is the history of the name, not of the bytes.