docs(registry): profile.md has no writer, and both surfaces that claimed one now say so (refs #58)

[atomchung]

What this fixes

Two surfaces asserted that ~/.trade-coach/profile.md is written by the agent per SKILL.md. Nothing writes it.

Location	The claim
docs/maintainer-guide.md:137	`profile.md` (the agent writes it directly per `SKILL.md`)
tests/test_coach_data_cli.py:303-304	profile.md (written directly by the agent per SKILL.md, never by engine code)

Verified against main@94b7e5f: grep -rn "profile" skills/fomo-kernel/SKILL.md skills/fomo-kernel/flows/ skills/fomo-kernel/references/ returns exactly two hits, both profile.json — a different file holding max_position_pct and muted_rules. No profile.md on any agent instruction surface, and no engine/ module writes it. It is a vestige of an earlier design that stayed registered.

Why both, and why together

git log -S puts the test comment first (6a82ec8, #471, when the registry-completeness scanner was written) and the guide's copy second (78aedc4, #587, when the shared contract left CLAUDE.md). The guide did not invent the claim — it inherited it.

Correcting only the guide would have left the assertion alive in the very file the guide's row cites as its coverage evidence: the mirrored-surface defect docs/maintainer-guide.md names as this repository's most frequent. Each corrected statement now names the other as its mirror, so a future edit to one is prompted to check the other.

The test comment also carried a second, smaller error the guide did not: it said both scan-invisible entries "are written from outside that scope", when only ux/ is. The two are now separated — they are unseen for different reasons, and only one of them has a writer.

What is deliberately unchanged

• coach.py's DATA_FILES. The entry stays registered, so data-status/data-export/data-reset keep seeing a file an older root may hold. Deregistering it would change export and reset coverage.
• Its user-facing description 「第一次復盤時建立,Claude 直接寫檔」 — the same stale claim in the one place a user reads it — and README.md:111 / README.zh-TW.md:111 / README.zh-CN.md:108, which still tell the user to cat it as "the baseline each review compares against".

Both are downstream of the revive / re-home / retire choice #58 owns, and are recorded there rather than fixed here. #597 separately ruled profile.md may not become a second canonical truth for provisional principles. This PR makes the maintainer authority stop testifying that a writer exists, so that choice gets made from measurement rather than from the guide.

Checked and left alone: tests/test_sizing_literal_gate.py:58 cites "the two paths it names as outside its own reach" as a scope-naming precedent. That statement stays true after this change, so it carries no drift.

Verification

python3 tests/run_all.py — 46/46 suites pass. tests/test_doc_language.py's maintainer-guide gate is presence-based ("a later edit to a row is that row's own business"), so a row rewrite is in scope for it.

Doc and comment text only; no behavior change, no schema change, no engine change.

Refs #58. Records the second location #597's evidence section flagged for re-homing.

Claim comment: #58 (comment)

🤖 Generated with Claude Code

[atomchung]

Blocking documentation correction before merge (could not submit REQUEST_CHANGES because the PR author and reviewer account are the same).

The measured defect and the two-surface correction are sound, but this branch now writes a stale owner into the shared maintainer authority.

docs/maintainer-guide.md says the revive / re-home / retire choice is #58's open call. Since this PR was opened, #619 was created specifically as the current decision owner, with #58 treated as the historical writer proposal and #446 as the structured successor. #27 and the bounded hygiene record on #538 now route the same way. Merging this wording would fix one false claim while installing a new stale execution pointer in the file maintainers are told to trust.

Please make the smallest correction:

• point the open decision to [decision] profile.md 是記憶體系第 2 層的化石:退役、復活,還是等 #446 刀 2 搬家? #619;
• describe feat: 第一次復盤後從行為推斷 profile.md + theses.jsonl（少問、猜為主） #58 as the historical proposal / evidence source rather than the current owner;
• keep [design·P1] The distillation layer: bounded profile memory over the append-only ledger (scoping) #446/[research·product] Philosophy bootstrap: decision guidance before transaction history #597 as the structural and truth-boundary references;
• update the PR body to the same ownership wording.

Also rebase/verify on latest main@9d6cb9a and report the current full offline suite rather than the pre-#611 46-suite count. No runtime widening is requested; coach.py, README and the actual retire/re-home decision remain out of this PR.