Answer prose round 1: split the record-scoring help into one idea per sentence, name the chunk's entries once, fold the changelog bullet into the bullets it changes, and give the decision's amendment its own paragraph

This commit is contained in:
2026-10-09 11:31:17 +02:00
parent c386163a5f
commit 6a8bbcfff9
9 changed files with 39 additions and 37 deletions
+1 -1
View File
@@ -12,7 +12,7 @@ Decisions live in `docs/decisions.md`:
- loop-claude: each project's model lives in `~/.claude/loop-claude-models.json`, which only loop-claude writes
- review: a pass gets at most three rounds per chunk; instead of a fourth, the run asks the human
- review: a fix for a 🔴 or 🟡 reruns the pass that found it and each pass whose trigger names fixes like it; a fix for a 🟢, a ❓ or a prose finding reruns none
- review: a paired ruling against the last measured commit decides the ratchet, and counts as worse when a dimension it rules fallen is under the floor; the four-seat scoring run checks the floor when a release's items run out, an item names the floor or a ruling names a fallen dimension
- review: a paired ruling against the last measured commit decides the ratchet, and counts as worse when a dimension it rules fallen is under the floor; the four-seat scoring run checks the floor when a release's items run out or an item names the floor, and also runs when a ruling names a fallen dimension
- review: a project decision may suspend the comprehension check for the plan items it names
- review: a pass skips only on a precedent: a decision, or a plan item crediting the pass, that the repo held before the chunk began
- loop-claude: the archive keeps the raw transcript, and a failed archive stops the loop
+2 -3
View File
@@ -2,10 +2,9 @@
## Unreleased
- larv-review's paired ruling takes each seat's vote on each dimension too: `Navigation same, Locality worse, Shape same, Self-sufficiency same, Overall better`. A better or same ruling in which a dimension fell calls a scoring run, and counts as worse, blocking the merge, when that dimension's mean is under the floor. Such a run blocks on the fallen dimensions alone and leaves the ratchet's base in place unless the ruling is better. Each place a seat names opens with the dimensions it moves, spelled as in the seat's line, such as `Locality, Shape: `, or the run fails; a dimension's vote counts only where one of the seat's places opens with it. `record-scoring` takes the shipped item IDs too and prints `Floor:` when the means under the floor block the merge.
- larv-review's mechanical rules run in the new, tested `bin/larv-review`: the paired ruling's base (`ruling-base`), each pass's round count (`round-count`), the scoring run's trigger (`needs-scoring-run <item ID>...`), and the vote count, means and history format (`record-ruling` and `record-scoring`, which read a seats file and append to `docs/comprehension-history.md`). The skill says when to run each and what its answer means. A seats file line the tool cannot read fails the run, so a mistyped vote never counts as same. The base search compares the history before and after each merged change, so it no longer depends on how a line is worded, and orders a change's rulings by their headings' times. A release whose last plan item moves to another release is due a scoring run too, once work has landed towards it. A seats file must name commits the checkout holds. Scores a `comprehension floor:` decision entry still holds move into the history through `record-scoring`, as a `Mean:` line. When origin cannot be fetched, the tool reads it as last fetched and says so.
- larv-review's mechanical rules run in the new, tested `bin/larv-review`: the paired ruling's base (`ruling-base`), each pass's round count (`round-count`), the scoring run's trigger (`needs-scoring-run <item ID>...`), and the vote count, means and history format (`record-ruling <seats file>` and `record-scoring <seats file> <item ID>...`, which read a seats file and append to `docs/comprehension-history.md`; `record-scoring` prints `Floor:` when the means under the floor block the merge). The skill says when to run each and what its answer means. A seats file line the tool cannot read fails the run, so a mistyped vote never counts as same. The base search compares the history before and after each merged change, so it no longer depends on how a line is worded, and orders a change's rulings by their headings' times. A release whose last plan item moves to another release is due a scoring run too, once work has landed towards it. A seats file must name commits the checkout holds. Scores a `comprehension floor:` decision entry still holds move into the history through `record-scoring`, as a `Mean:` line. When origin cannot be fetched, the tool reads it as last fetched and says so.
- The new `bin/metric-bench` benches cheap static metrics of a project's Python against the places its comprehension history names as hardest: cognitive complexity, length, shared state and change spread, each scored by how well it ranks the named functions above the rest. No metric enters review yet.
- larv-review blocks a merge when simulated readers rule that the PR makes the project read worse than the last measured commit: the default branch's newest merged change whose last ruling in `docs/comprehension-history.md` is better, or that added a scoring run there while that ruling named no fall. A ruling of same leaves that commit in place, so a worse ruling may name places earlier chunks wrote. In each architecture pass Mid A and the inherited architect vote better, same or worse; when they disagree, the maintainability senior and Junior A vote too. A ruling needs a lead of two or more votes, else it is same, and a vote that names no place counts as same. A worse ruling never ships: after three rounds larv-review asks you for another round, a no leaves the PR unmerged, and so does an unattended run. The four-seat scoring run checks the comprehension floor only when a chunk ships or removes a release's last plan item, or ships an item naming the floor; between those runs the floor blocks only a dimension a paired ruling finds fallen. A score under the floor blocks the merge until fixed or until you decide to ship; a decision in `docs/decisions.md` that lets every chunk ship under the floor counts as yours.
- larv-review blocks a merge when simulated readers rule that the PR makes the project read worse than the last measured commit: the default branch's newest merged change whose last ruling in `docs/comprehension-history.md` is better, or that added a scoring run there while that ruling named no fall. A ruling of same leaves that commit in place, so a worse ruling may name places earlier chunks wrote. In each architecture pass Mid A and the inherited architect vote better, same or worse on each dimension and overall: `Navigation same, Locality worse, Shape same, Self-sufficiency same, Overall better`. When their overall votes differ, the maintainability senior and Junior A vote too. A ruling needs a lead of two or more votes, else it is same. Each place a seat names opens with the dimensions it moves, such as `Locality, Shape: `, or the run fails, and a vote that no place opens with counts as same. A dimension falls when its worse votes lead by two. A better or same ruling in which a dimension fell calls a scoring run, and counts as worse when that dimension's mean is under the floor; such a run leaves the ratchet's base in place unless the ruling is better. A worse ruling never ships: after three rounds larv-review asks you for another round, a no leaves the PR unmerged, and so does an unattended run. The four-seat scoring run checks the comprehension floor only when a chunk ships or removes a release's last plan item, or ships an item naming the floor; between those runs, a score under the floor blocks a merge only through a fallen dimension, and that block counts as a worse ruling. On those runs, a score under the floor blocks the merge until fixed or until you decide to ship; a decision in `docs/decisions.md` that lets every chunk ship under the floor counts as yours.
- larv-review reruns fewer passes after a fix. A fix reruns the pass whose 🔴 or 🟡 it answers, plus stability when the fix changes behaviour, security when it opens an untrusted-input boundary or touches authn/authz, secrets or crypto, and architecture when it carries an architectural decision. A rerun reviews only what changed since that pass last reviewed. A fix for a 🟢, a ❓ or any prose finding reruns no pass, and prose takes one round.
- larv-review runs one security pass on each change: the full scan alone when it runs, else `/security-review`.
- larv-review runs the new `bin/reading-cost`, which measures each chunk `loop-claude` finished from its archive: the new input tokens the worker and its subagents took in before the worker's first commit, and the files git tracks that they read most. It appends each figure the project's `docs/reading-cost.md` lacks. Once 10 chunks are measured, and the last 5 chunks' median is above the 5 before them, larv-review files one plan item naming the files read most, or refreshes it. The chunk that ships it runs `reading-cost --reset`, so the trend counts only chunks that finished after it.
+1 -1
View File
@@ -115,7 +115,7 @@ def appended(history: str, entry: str) -> str:
def ruling_entry(title: str, seats: list[Seat], said: dict[str, list[str]], counted: dict[str, list[str]], ruled: str, fell: list[str]) -> str:
"""A paired ruling's entry, with a `Fell:` line when fell names dimensions; a seat's vote that counted as same for want of a place says so."""
"""A paired ruling's entry; a seat's vote that counted as same for want of a place says so."""
votes = [[seat.name, *map(vote_text, DIMENSIONS, said[seat.name], counted[seat.name])] for seat in seats]
lines = [f"Ruling: {ruled}"] + ([f"Fell: {', '.join(fell)}"] if fell else [])
return entry_text(title, lines, [["Seat", *DIMENSIONS], *votes], [(f"{seat.name}, {DECIDED_BY}:", seat.places) for seat in seats])
+19 -18
View File
@@ -1,6 +1,6 @@
"""Hold larv-review's mechanical rules: the paired ruling's base and vote count, each pass's rounds, the scoring run's trigger, and the comprehension history's format.
usage: larv-review ruling-base | round-count | needs-scoring-run [<item ID>...] | record-ruling <seats file> | record-scoring <seats file>
usage: larv-review ruling-base | round-count | needs-scoring-run [<item ID>...] | record-ruling <seats file> | record-scoring <seats file> [<item ID>...]
Run it inside a project checkout or one of its worktrees. Each command but record-ruling first
fetches origin, when the checkout has that remote, and takes origin's default branch, or else the
@@ -19,10 +19,11 @@ round-count prints each pass's count: the highest <n> among the titles opening
cap of {cap} rounds, and each name in such a title that is not a pass.
needs-scoring-run takes the IDs of the todo.md items the chunk ships, none when it ships none, and
prints why the chunk is due a scoring run, or `No scoring run due.` It is due when the chunk ships
an item whose tagline names the comprehension floor, when the checkout's history ends on a paired
ruling with a `Fell:` line, or when it leaves a release with no item in
the checkout's todo.md, less the shipped items, and work has landed towards that release. Work has
prints why the chunk is due a scoring run, or `No scoring run due.` The chunk's entries are those
the checkout's history holds and the default branch's lacks. It is due when the chunk ships an
item whose tagline names the comprehension floor, when the chunk's newest entry is a paired ruling
with a `Fell:` line, or when it leaves a release with no item in the checkout's todo.md, less the
shipped items, and work has landed towards that release. Work has
landed when the chunk ships one of the release's items, or when the release is the earliest the
default branch's todo.md names and the default branch has a commit after its newest v* tag, or has
no v* tag.
@@ -44,11 +45,12 @@ opens with its label, written exactly: `Junior A:`, `Mid A:`, `Maintainability s
- Locality: `bin/x.py:12` `run`: ...
record-ruling takes Mid A's and the inherited architect's votes on each dimension and overall,
each better, same or worse. Each place opens with the dimensions it moves, or Overall, spelled as
the seat's line spells them and joined by `, `, then `: `. A dimension's vote counts as same unless one of the seat's places in
backticks, `path:line`, names that dimension; the overall vote counts as same unless the seat names
any such place. The printed votes say so. When the two counted overall votes differ, it appends
nothing and prints `Split:`, saying to add the maintainability senior's and Junior A's lines.
each better, same or worse. Each place opens with the names from the seat's line that it moves,
joined by `, ` and followed by `: `. A dimension's vote counts as same unless one of the seat's
places opens with that dimension and holds a `path:line` in backticks; the overall vote counts as
same unless any place holds one. The printed votes say so. When the two counted overall votes
differ, it appends nothing and prints `Split:`, saying to add the maintainability senior's and
Junior A's lines.
Otherwise the ruling is the overall vote that leads each other vote by two or more, else same; it
prints `Ruling:` and appends the ruling. A dimension falls when its worse votes lead each other
vote by two or more. Unless the ruling is worse, it prints and records a `Fell:` line naming each.
@@ -60,13 +62,12 @@ Each line is `Navigation <n>, Locality <n>, Shape <n>, Self-sufficiency <n>, Ove
score from 1 to 10 and the overall at most 1 above the line's lowest dimension. Its first line
names no `, against`. It appends the seats' scores and their means, rounded half up to two
decimals, and prints the means and each one under the floor of {floor}. It exits 1 on a run with
seats that no plan item, release or fall calls. A `Floor:` line says the
means under the floor block the merge, as they do when needs-scoring-run would give a reason
besides a fall. When the history ends on a
paired ruling with a `Fell:` line, it says whether the ruling stands; when a dimension that fell is
under the floor, the ruling counts as worse, and it prints and records a `Ratchet:` line, which
needs-scoring-run repeats as `Still standing:` until a paired ruling follows it. Both read only the entries the default
branch's history lacks.
seats that no plan item, release or fall calls. A `Floor:` line says the means under the floor
block the merge, as they do when needs-scoring-run would give a reason besides a fall. It reads the
chunk's entries as needs-scoring-run does. When a fall is pending, it says whether the paired
ruling stands. When a fallen dimension's mean is under the floor, the ruling counts as worse, and
it prints and records a `Ratchet:` line. needs-scoring-run repeats that line as `Still standing:`
until a paired ruling follows it.
Each record is appended to the history, created when missing, under a heading of the current UTC
time and the seats file's first line. A file holding a `Mean:` line alone opens with the date or
@@ -226,7 +227,7 @@ def record_ruling(seats_file: pathlib.Path) -> Failure | None:
fell = fallen(count) if ruled != "worse" else []
print(f"Ruling: {ruled}. Overall votes as counted: {tally}.")
if fell:
print(f"Fell: {', '.join(fell)}. A scoring run is due; the ruling counts as worse if it puts {' or '.join(fell)} under the floor of {FLOOR}.")
print(f"Fell: {', '.join(fell)}. A scoring run is due; the ruling counts as worse if that run scores {' or '.join(fell)} under the floor of {FLOOR}.")
return append(ruling_entry(title, given.seats, said, count, ruled, fell))
+5 -5
View File
@@ -33,7 +33,7 @@ def votes(seats: list[Seat]) -> dict[str, list[str]] | Failure:
def counted(seats: list[Seat], said: dict[str, list[str]]) -> dict[str, list[str]]:
"""Each seat's votes as they count: a dimension's vote counts as same unless one of the seat's `path:line` places names that dimension, and the overall vote unless the seat names any such place."""
"""Each seat's votes as they count: a dimension's vote counts as same unless one of the seat's `path:line` places opens with that dimension, and the overall vote unless the seat names any such place."""
return {seat.name: [vote if names_a_place(seat, name) else "same" for name, vote in zip(DIMENSIONS, said[seat.name])] for seat in seats}
@@ -63,13 +63,13 @@ def fallen(counted: dict[str, list[str]]) -> list[str]:
def chunk_entries(default_history: str, history: str) -> list[Entry]:
"""The checkout's entries whose headings the default branch's history lacks, oldest first; of entries with one time, the later in the file is newer."""
"""The checkout's entries whose headings the default branch's history lacks, oldest first; entries sharing a time keep their file order."""
held = {entry.heading for entry in entries(default_history)}
return sorted((entry for entry in entries(history) if entry.heading not in held), key=lambda entry: entry.time)
def pending_fall(chunk: list[Entry]) -> list[str]:
"""The dimensions that fell in the chunk's newest entry: a paired ruling no scoring run has followed yet."""
"""The dimensions that fell in the chunk's newest entry, empty unless that entry is a paired ruling."""
return chunk[-1].fell if chunk else []
@@ -89,7 +89,7 @@ def floor_verdict(under: list[str], reasons: list[str]) -> str:
return f"None under the floor of {FLOOR}."
if reasons:
return f"Floor: {', '.join(under)} under the floor of {FLOOR} block the merge, as the chunk {' and '.join(reasons)}."
return f"Under the floor of {FLOOR}, blocking nothing on a run due for no reason besides a fall: {', '.join(under)}."
return f"Under the floor of {FLOOR}, blocking nothing, as only a fall called this run: {', '.join(under)}."
def fall_outcome(fell: list[str], found: list[decimal.Decimal]) -> tuple[str, list[str]] | None:
@@ -99,7 +99,7 @@ def fall_outcome(fell: list[str], found: list[decimal.Decimal]) -> tuple[str, li
if dropped := [f"{name} is {value}" for name, value in zip(DIMENSIONS, found) if name in fell and value < FLOOR]:
line = f"Ratchet: the paired ruling counts as worse, as {' and '.join(fell)} fell and {', '.join(dropped)}, under the floor of {FLOOR}."
return line, [line]
return f"The paired ruling stands: {' and '.join(fell)} fell and stays at or above the floor of {FLOOR}.", []
return f"The paired ruling stands: {' and '.join(fell)} fell and each stays at or above the floor of {FLOOR}.", []
def means(lines: list[list[str]]) -> list[decimal.Decimal]:
+3 -3
View File
@@ -80,7 +80,7 @@ class Seats(unittest.TestCase):
found = [decimal.Decimal(score) for score in ("6", "6.5", "7", "6", "6")]
ratchet = "Ratchet: the paired ruling counts as worse, as Locality and Shape fell and Locality is 6.5, under the floor of 7.0."
self.assertEqual(fall_outcome(["Locality", "Shape"], found), (ratchet, [ratchet]))
self.assertEqual(fall_outcome(["Shape"], found), ("The paired ruling stands: Shape fell and stays at or above the floor of 7.0.", []))
self.assertEqual(fall_outcome(["Shape"], found), ("The paired ruling stands: Shape fell and each stays at or above the floor of 7.0.", []))
self.assertIsNone(fall_outcome([], found))
def test_rules_by_a_lead_of_two_and_splits_two_seats_that_differ(self):
@@ -315,14 +315,14 @@ class LarvReviewTool(unittest.TestCase):
seats = "".join(f"\n{seat}: {line}\n" for seat in ("Junior A", "Mid A", "Maintainability senior", "Inherited architect"))
scored = self.run_tool("record-scoring", self.seats_file(f"PR #2 at {self.start}\n" + seats))
self.assertIn("Ratchet: the paired ruling counts as worse, as Locality fell and Locality is 6.50, under the floor of 7.0.", scored.stdout)
self.assertIn("blocking nothing on a run due for no reason besides a fall: Locality 6.50.", scored.stdout)
self.assertIn("blocking nothing, as only a fall called this run: Locality 6.50.", scored.stdout)
self.assertIn("\n\nRatchet: the paired ruling counts as worse", (self.checkout / "docs" / "comprehension-history.md").read_text())
self.assertTrue(self.run_tool("needs-scoring-run", "2").stdout.startswith("No scoring run due.\nStill standing: Ratchet:"))
same = self.run_tool("record-ruling", self.seats_file(f"PR #3 at {self.start}, against {self.start}\n\nMid A: {said('same', 'worse')}{place}Inherited architect: {said('same', 'worse')}{place}"))
self.assertIn("Fell: Locality.", same.stdout, same.stderr)
stands = self.run_tool("record-scoring", self.seats_file(f"PR #2 at {self.start}\n" + seats.replace("Locality 6.5", "Locality 7")))
self.assertIn("The paired ruling stands: Locality fell and stays at or above the floor of 7.0.", stands.stdout)
self.assertIn("The paired ruling stands: Locality fell and each stays at or above the floor of 7.0.", stands.stdout)
def test_names_no_fall_on_a_worse_ruling(self):
place = "\n\n- Locality, Overall: `bin/a.py:3` `f`: longer\n\n"
+4 -2
View File
@@ -65,9 +65,11 @@ lilleman decided to ship each of these under the floor on the day listed. Each l
2026-10-08, lilleman, replacing the 2026-09-30 rule that every fix, a nit's included, re-enters review. Serves goal 1.4. An audit of the four costliest archived chunks (2026-10-08) found nearly every finding that mattered in a pass's first round; later rounds mostly re-checked the last fix, reworded what an earlier round wrote, or contradicted another round on the same line. A round 2 caught real bugs only where a fix changed behaviour, so stability still reruns on every fix that does, and security on every fix that opens an untrusted-input boundary or touches authn/authz, secrets or crypto. Valid while round-2 reviews of nit and prose fixes keep changing nothing a user would notice.
## review: a paired ruling against the last measured commit decides the ratchet, and counts as worse when a dimension it rules fallen is under the floor; the four-seat scoring run checks the floor when a release's items run out, an item names the floor or a ruling names a fallen dimension
## review: a paired ruling against the last measured commit decides the ratchet, and counts as worse when a dimension it rules fallen is under the floor; the four-seat scoring run checks the floor when a release's items run out or an item names the floor, and also runs when a ruling names a fallen dimension
2026-10-08, lilleman. Serves goals 1.1 and 1.4. Two independent four-seat runs moved up to 0.4 on near-identical code, so a fall within that noise forced rounds that changed nothing (PR #224). A seat reading both versions judges the difference on the same material, so the noise the versions share cancels. No conventional significance level lets a sign test settle with four votes, so the human chose a lead of two votes over every other ruling, same included. Two seats that agree settle it, so a PR that leaves the reading as it was costs two seats. A third seat alone can never settle it, so the third and fourth launch together. The base is the default branch's newest merged change whose last added paired ruling is better, or that adds a scoring run to the history while that ruling names no fall. The human chose the last measured commit on the architect's finding that with a merge base as the base, chunks without an architecture pass, and a suspended restructure's steps, could each lower comprehension unseen. A ruling of same leaves the base where it was, or a run of small falls each ruled same would add up unseen. A scoring run moves it too: otherwise a project never ruled better would compare against its first ruling forever, and a release's floor check is where the human sees the absolute level. Reading the default branch one merged change at a time keeps the rule working for merge commits and squash merges alike: a branch's own early ruling of better never counts, only the merged change's last ruling. The scoring run also runs on a chunk shipping an item that names the floor, so the chunk that lifts a project to the floor can prove it. A chunk that removes a release's last item without shipping it runs it too, so no release ends unmeasured. 2026-10-09, lilleman: each seat also rules each dimension, and a dimension falls by the same lead of two. A better or same ruling naming a fallen dimension calls the scoring run, and counts as worse when that dimension's mean is under the floor, so between releases a project cannot trade one dimension below the floor for gains in the other three. Such a run blocks on the fallen dimensions alone: the floor's own block stays with a release's end and an item naming the floor. Nor does it move the base, unless the ruling is better: a run of falls each ruled same must not add up unseen. A dimension's vote counts only where one of the seat's places names that dimension. Valid while seats reading both versions agree more often than two absolute runs do.
2026-10-08, lilleman. Serves goals 1.1 and 1.4. Two independent four-seat runs moved up to 0.4 on near-identical code, so a fall within that noise forced rounds that changed nothing (PR #224). A seat reading both versions judges the difference on the same material, so the noise the versions share cancels. No conventional significance level lets a sign test settle with four votes, so the human chose a lead of two votes over every other ruling, same included. Two seats that agree settle it, so a PR that leaves the reading as it was costs two seats. A third seat alone can never settle it, so the third and fourth launch together. The base is the default branch's newest merged change whose last added paired ruling is better, or that adds a scoring run to the history while that ruling names no fall. The human chose the last measured commit on the architect's finding that with a merge base as the base, chunks without an architecture pass, and a suspended restructure's steps, could each lower comprehension unseen. A ruling of same leaves the base where it was, or a run of small falls each ruled same would add up unseen. A scoring run moves it too: otherwise a project never ruled better would compare against its first ruling forever, and a release's floor check is where the human sees the absolute level. Reading the default branch one merged change at a time keeps the rule working for merge commits and squash merges alike: a branch's own early ruling of better never counts, only the merged change's last ruling. The scoring run also runs on a chunk shipping an item that names the floor, so the chunk that lifts a project to the floor can prove it. A chunk that removes a release's last item without shipping it runs it too, so no release ends unmeasured. Valid while seats reading both versions agree more often than two absolute runs do.
2026-10-09, lilleman: each seat also rules each dimension, and a dimension falls by the same lead of two. A better or same ruling naming a fallen dimension calls the scoring run, and counts as worse when that dimension's mean is under the floor, so between releases a project cannot trade one dimension below the floor for gains in the other three. Such a run blocks on the fallen dimensions alone: the floor's own block stays with a release's end and an item naming the floor. Nor does it move the base, unless the ruling is better: a run of falls each ruled same must not add up unseen. A dimension's vote counts only where one of the seat's places opens with that dimension, so a fall rests on places the seats name.
## review: a project decision may suspend the comprehension check for the plan items it names
+1 -1
View File
@@ -90,5 +90,5 @@ Same means a reader of either version would find it as easy to understand and ch
Report, in this shape:
1. Your rulings, each better, same or worse, in this shape: Navigation <ruling>, Locality <ruling>, Shape <ruling>, Self-sufficiency <ruling>, Overall <ruling>
2. The places that decided them, most decisive first. Each opens with the dimensions it moves, spelled as in your rulings line, or Overall, joined by ", " and followed by ": ". Then path:line and the unit's name, in the version the place is in, and one or two sentences on what reads easier or harder there than in the other version. Rulings with no named place are discarded.
2. The places that decided them, most decisive first. Each opens with the names from your rulings line that it moves, joined by ", " and followed by ": ", such as `Locality, Shape: `. Then `path:line` and the unit's name, each in backticks, in the version the place is in, and one or two sentences on what reads easier or harder there than in the other version. A ruling that no place opens with counts as same.
```
+3 -3
View File
@@ -101,12 +101,12 @@ Then the comprehension check: simulated readers, launched per `~/.claude/skills/
The chunk does not merge while either holds:
- **Ratchet.** The paired ruling is worse, or `record-scoring` prints `Ratchet:` because a dimension that fell is under the floor, which `needs-scoring-run` repeats as `Still standing:` until a paired ruling follows it.
- **Floor.** `record-scoring` prints `Floor:`, naming means under the floor of 7.0 on a run due for a reason besides a fall.
- **Ratchet.** The paired ruling is worse, or `record-scoring` prints `Ratchet:`, because a dimension that fell is under the floor. `needs-scoring-run` repeats that line as `Still standing:` until a paired ruling follows it.
- **Floor.** `record-scoring` prints `Floor:`, naming means under the floor of 7.0.
Each is a 🔴 of the comprehension check. It is never filed, and stands until a later run clears it or, for the floor, a decision to ship answers it.
Fix a worse ruling at the places the seats voting worse name, a fallen dimension at the places they name under it, and the floor at the places the seats name under the lowest dimension. Under the floor alone, the user may decide to ship instead. A decision that lets every chunk ship under the floor, held in the project's `docs/decisions.md` before the chunk began, counts as theirs while it stands. A decision to ship listed in the `comprehension floor:` entry covers only the chunk it names. No user reachable → fix. A worse ruling never ships: after three rounds, "Round cap" → "Cap" asks the user for another round.
Fix a worse ruling at the places the seats voting worse name, a fallen dimension at the places that open with it, and the floor at the places the seats name under the lowest dimension. When only the Floor bullet holds, the user may decide to ship instead. A decision that lets every chunk ship under the floor, held in the project's `docs/decisions.md` before the chunk began, counts as theirs while it stands. A decision to ship listed in the `comprehension floor:` entry covers only the chunk it names. No user reachable → fix. A worse ruling never ships: after three rounds, "Round cap" → "Cap" asks the user for another round.
Two files of the project keep the records: