fix(ci): stop the review bot handing over fixes in prose

The `suggestion` blocks stopped once the briefing moved into its own file, but
the carve-out that survived — "one clause naming where the fix belongs" — was
being stretched from a location into an instruction. #6397 dictated what to
write in a comment and which existing test to copy; #6394 named the fix
outright. The clause now permits a file, a function, a symbol or a layer and
nothing about what happens there, and closes the stretch three ways: prose is
a patch the moment a verb describes the change, so is holding up an existing
symbol as the model to copy, and a clause the maintainer could apply as
written is the fix however it is punctuated.

Three rules the rubric was missing, none of which existed anywhere. A 🔴 or 🟡
says in one clause what the change did to the code it is about, the way a 🟣
already says it predates it — otherwise nothing in the comment shows the
marker was earned. A claim about a caller or a callee needs that file read:
the dispatch-rule violation this repo cares most about sits a frame outside
the diff, and the skill is told to avoid reading past the changes. And nothing
pads the comment.

The briefing's one named override aimed at a step that does not exist. The
plugin the job loads defines no `--comment` flag and mentions suggestions
nowhere, so `max --comment <target>` is inert trailing text. Replaced with the
six overrides that are real: the skill calls pre-existing issues and unmodified
lines false positives, drops every finding its confidence pass scores under 80
and then posts nothing at all (a nitpick scores 50, so that filter empties all
five nit slots), says to avoid emojis against a severity system that is three
of them, mandates a "Found N issues" format, and forbids reading build signal.

.github/claude/review-job.md

@@ -15,15 +15,31 @@ report, the repo-specific checks, the verification bar and the volume cap. The
 skill loads `CLAUDE.md` on its own; it does not load `REVIEW.md`, which is why
 this briefing exists.
 
+Where the skill and `REVIEW.md` disagree, `REVIEW.md` wins. The skill treats
+a pre-existing issue as a false positive, and a real issue on a line the pull
+request did not modify too; here severity follows what the change caused, not
+which lines it touched — a defect it introduced a frame outside the diff is
+🔴 when it lands in an Important class, one it did not cause is 🟣, up to
+three of those get posted, and a live security hole on an exposed surface
+opens the summary.
+
+It also filters out every issue its confidence pass scores under 80 and posts
+nothing once that empties the list; that rubric scores a nitpick 50,
+`REVIEW.md` allots five nits, and the comment goes up either way. It says to
+avoid emojis, and the whole severity system is three of them. Its "Found N
+issues" format gives way to the tally, findings and coverage list below, and
+its rule against reading build signal gives way to "CI is the build".
+
 ## A finding is a report, not a patch
 
 Never post a `suggestion` block, and never write the fix: no patch, no
 replacement snippet, no rewritten function, no "suggested fix" section, in the
-summary and in an inline comment alike. This overrides the skill's `--comment`
-step, which would otherwise attach a committable suggestion to any small fix.
-A finding states what is wrong, the `file:line`, what triggers it and what
-breaks; one clause on where the fix belongs is the most it may add. The
-maintainer decides the change.
+summary and in an inline comment alike. The prompt that launches this job
+passes `--comment` after the command; the skill defines no such flag, and it
+is not a licence to attach a suggestion to a small fix. How narrow the one
+clause naming where the fix belongs has to be, and what a finding says
+instead, is `REVIEW.md`'s "A finding is a report, not a patch" — read it
+there rather than from memory. The maintainer decides the change.
 
 ## Skip gate
 

REVIEW.md

@@ -103,6 +103,10 @@ near-certain about and that actually breaks something:
 
 - A claim about behaviour needs a `file:line` citation from this repository,
   not an inference from a name.
+- A claim about what the change does to a caller or a callee needs that file
+  read, not inferred from the hunk. A dispatch-rule violation rarely shows
+  inside the diff — the changed line calls an innocuous helper and the
+  `internal/xray/api.go` call sits a frame outside it.
 - A claim that a downstream client rejects or requires a wire-format detail —
   a config key, JSON tag, URI query parameter, YAML or TOML key, an encoding
   or hash choice — must name the upstream symbol that decides it (repository,
@@ -137,6 +141,10 @@ Open with a one-line tally — `2 🔴 / 4 🟡 / 1 🟣` — so the author sees
 shape of the review before the detail. When nothing is 🔴, lead with
 `No blocking issues` and put the tally after it.
 
+Nothing pads the comment: no "Strengths" section, no restatement of what the
+pull request does, no praise, no closing pleasantry. Padding is not neutral —
+it buries the two lines someone actually has to act on.
+
 The posted comment is the only part of a review anyone sees, so a bare "no
 issues found" is a receipt, not a review: nothing in it says whether the diff
 was read or the run died early. Every comment therefore ends with a short
@@ -151,6 +159,18 @@ evidence, not a retelling of the pull request.
 A finding says what is wrong, where (`file:line`), what triggers it and what
 breaks. It never carries the fix: no `suggestion` block, no patch, no
 replacement snippet, no rewritten function, no "suggested fix" section — in
-the summary and in an inline comment alike. One clause naming where the fix
-belongs is the most it may add. The maintainer decides the change; a review
-that writes it out puts unreviewed code one click from the branch.
+the summary and in an inline comment alike. One clause naming WHERE the fix
+belongs is the most it may add — a file, a function, a symbol, a layer — and
+nothing about what happens there. Prose is a patch too the moment a verb
+describes the change: "move the lookup inside the body", "spend the comment
+on the invariant instead" hand it over as surely as a diff would, and so does
+holding up an existing symbol as the model to copy. A clause the maintainer
+could apply as written is the fix, however it is punctuated. The maintainer
+decides the change; a review that writes it out puts unreviewed code one
+click from the branch.
+
+A 🔴 or 🟡 finding also says, in one clause, what this pull request did to
+the code it is about — the line it added, the call it moved, the guard it
+dropped — the way a 🟣 says that it predates the change. That clause reports
+what the change did, never what it should have done. Nothing else in the
+comment shows the marker was earned.