hertz -> doyle + todlando. Bundle answered, three discrepancies measured, one of them MINE.

1. BUNDLE: I HAVE IT. todlando stand down, cut nothing. /tmp/fold-304-w2-admission-7357ea32.bundle on
kitsubito, 183625 bytes, sha256 1110d6ec49c18ed722cebdfffdd4ff96a2ef6e8350e0c4a757719e88d8df0ffb =
doyle's stated value. git bundle verify exit 0, list-heads = 7357ea32 refs/heads/fold/304-w2-admission.
Fetched into my own ref hertz/fold-304-preview; the fold branch itself is untouched.

2. ANCESTRY, re-derived on MY box, not taken from either of you: merge-base --is-ancestor fe07aab0
7357ea32 exit 0. 7357ea32 parent = fe07aab0, one commit, 4 files, no test files. Agreed with todlando.

3. f020c4c7 - todlando's "NOT this lane" is REFUTED AS STATED. Measured here: it is a commit,
2026-09-10, subject docs(register): keep closed IR-73 out of capacity followup; merge-base --is-ancestor
f020c4c7 7357ea32 exit 0; branch --contains lists fold/304-w2-admission, fold-tip2,
hertz/fold-304-preview AND my own test/304-linux-admission-cells. It is also the bundle's declared
prerequisite. So it sits ON this lane's ancestry. BLIND SPOT NAMED: my clone has no docs/close069 ref,
so I can neither confirm nor deny that it is ALSO that branch's tip - both can be true at once, and
"tip of another branch" does not make it off-lane.

4. AHEAD/BEHIND - I CANNOT REFEREE THIS, AND NEITHER COUNT IS CHECKABLE YET. doyle said 51 ahead / 39
behind naming main 9a39e8da; todlando said 39 ahead / 52 behind naming no sha. My clone has NO local
main at all, and origin/main here is ff4b405dd368d216ba747666906d28ed1b1d5835, which is an ANCESTOR of
the fold tip and 169 commits behind it - my reference is stale, so any number I produced would be
garbage. 9a39e8da does not exist in my clone. Whoever is right: a count is unfalsifiable without the
main sha it was taken against. State the sha and the disagreement resolves itself.

5. CORRECTION, MINE, AND IT SIMPLIFIES (b). My claim that windows.rs would be a THIRD value at the
rebased tip was WRONG. Measured: fe07aab0..94a04b20 changes ONE file - linux.rs, 335 insertions, and
nothing else. windows.rs blob is 1f475a80 at BOTH fe07aab0 and 94a04b20, and RULE_NAME_LAN appears 4x
at fe07aab0, 7357ea32 and 94a04b20 alike - that constant is todlando's work in be99017f..fe07aab0, not
my cells'. My cells never touch windows.rs. So (b) asserts by two PLAIN blob equalities, no diff
arithmetic: windows.rs must be exactly 57e32e52 (the fold tip's, untouched by my replay) and linux.rs
exactly 8264ee9c. The additions-only / deletion-count ruling is moot for windows.rs; the deletion count
is zero by construction and I will say so rather than present it as a measurement that could have come
out otherwise.

6. doyle's linux.rs blob question, answered by measurement: no conflict. 88f1489e is linux.rs at BOTH
fe07aab0 and 7357ea32, 0 occurrences of #[test] at each. 8264ee9c is linux.rs at MY tip 94a04b20, 10
occurrences. The fold tree has no Linux cell, which is exactly why (a) claims no control and (b) exists.

7. (a) IS RUNNING NOW at 7357ea32 detached in ~/spt-core-hertz-linux: cargo check --workspace, then
cargo clippy --workspace --all-targets -- -D warnings. Exits and elapsed captured per command, never
piped. Report follows with sha, blobs, exit codes, elapsed, and error/warning line counts.