diff --git a/docs/remediation/TASKS.md b/docs/remediation/TASKS.md
index 198e146b..5cb653bc 100644
--- a/docs/remediation/TASKS.md
+++ b/docs/remediation/TASKS.md
@@ -736,6 +736,64 @@ that survives a counterexample the alternative fails is better endorsed than one
 lose that reason** — if the squash takes the merger, rotation protects nothing. Decisions stand on their
 other reasons; **the reason that _felt_ decisive is the one that was not measured.**
 
+### D-55d — I MEASURED both fields, ANALYSED one, and thereby hid the answer for an hour
+
+**The rule resolves on TWO fields: `author` = POSTER, `committer` = MERGER. Both roles are recorded, in
+different fields.** Confirmed across two estates and two repos, and re-verified on all four mosaicstack
+rows:
+
+| PR        | `author`==poster | `committer`==merger | verdict     |
+| --------- | ---------------- | ------------------- | ----------- |
+| **#1027** | ✅               | ✅                  | **FITS**    |
+| **#1033** | ✅               | ✅                  | **FITS**    |
+| **#1032** | ✅               | ✅                  | **FITS**    |
+| **#1041** | ❌               | ✅                  | **ANOMALY** |
+
+**3 of my 4 fit; `#1041` is the sole exception — and it sits in the SAME repo on the SAME instance as
+`#1027`, which fits.** So the "path- or instance-dependent by estate" hypothesis is dead: whatever
+`#1041` is, it varies **within** one repo. **No mechanism is proposed — one anomalous row does not earn
+one.**
+
+> **★ MY ERROR, AND IT IS A THIRD DISTINCT VARIANT OF TONIGHT'S CLASS. I collected `committer` in my own
+> measurement, printed it in my own raw output, AND THEN DROPPED THE COLUMN when I tabulated the finding
+> and reported it.** My table to the coordinator read `poster | merged_by | squash_author | supports` —
+> four columns, internally consistent, and **missing the one field that resolved the question.** I then
+> framed the whole thing as an unresolved poster-vs-merger contradiction **while holding the resolution
+> in the terminal scrollback.**
+
+**Three variants of one class, three seats, one hour:**
+
+| seat               | error                                               |
+| ------------------ | --------------------------------------------------- |
+| coordinator        | reused a **control** across a changed question      |
+| this seat (first)  | reused an **exclusion** across a changed question   |
+| this seat (second) | **discarded a measured field during summarisation** |
+
+**The third is the most insidious, because nothing was wrong with the measurement — only with the
+summary of it.** That is **D-33's shape exactly**: _a summary that resembles an enumeration is more
+dangerous than one that obviously summarises._ A four-column table **looks** like the data.
+
+> **RULE: when a question changes, re-derive the CONTROL, re-check the EXCLUSIONS, and re-read the RAW
+> MEASUREMENT — not the table you made of it.** The table was built to answer the previous question.
+
+**Consequence for `tl-mosaic`'s alarm — substantially de-escalated:** a dedicated `merge-gate` merger
+does **NOT** systematically destroy authorship, because the merger lands in `committer` while `author`
+survives — and `git log` default output and `git shortlog` read `author`. `#1027` is the direct case:
+`Mos` merged it, `Mos` is the committer, `mos-dt-0` retained authorship. **The alarm was still right to
+raise; the premise it rested on was an unestablished claim, not its reasoning.**
+
+**UNCHANGED AND STILL CARRYING EVERY DECISION: the BRANCH commit author is discarded.** Trailers remain
+required.
+
+> **★ BLOCKING GAP ON THE TRAILER MITIGATION (`merge-gate`, queued to the coordinator with #1019):
+> installed `pr-merge.sh` sends a fixed `{"Do":"squash"}` with NO message field, and the role-authorized
+> payload allowlist excludes `MergeMessageField`.** **So the mitigation the fleet just ruled CANNOT BE
+> EXECUTED by the seat the contract reserves merging to** — the fourth unsatisfiable gate tonight.
+> Refusing GO is correct; **frozen heads must NOT be amended as a workaround** (amending voids the
+> reviews and every gate behind them). **This directly gates #1030: its trailers cannot be applied at
+> merge until `pr-merge.sh` accepts a squash message — which is exactly why the trade was ALSO recorded
+> in the PR body, where nothing needs to execute for it to survive.**
+
 ### D-55b — two parties quoting a third is not corroboration
 
 The coordinator wrote _"installer-7's live-panel GREEN was the last thing before the command"_ **having