@codex-reviewer-andresmgsl's three findings (#4712), each a route by which
an unprovable read could still be reported as a whole one — the guard
leaking the failure class it exists to stop.
1. x-total-count was never validated. `X-Total-Count: not-a-number` returned
rc=0 with that string as the bound the walk compared against, reproduced
on ab23a3b. Now required to be a canonical non-negative integer.
2. The total was read once. A collection changing size under the walk was
invisible: page 1 declaring 4 and page 2 declaring 9 stopped at 4
believing itself whole. Now re-read per page; a moving total means the
read was not atomic and is refused.
3. A 200 whose body is not an array counted as zero items, so an error
object or scalar arriving where a list belongs read as a complete EMPTY
collection whenever the declared total was 0. Now refused, quoting the
body. A genuinely empty array is still fine — covered.
Each guard is mutation-verified: removing it reds exactly its own cases and
no others.
Refs #188
SC2016 on the deliberate single-quoted bash -c (the expansion belongs to
the isolated process, as the sibling case in issueflow-reconcile.test.sh
already documents), and SC2317 on the curl stub, which shellcheck cannot
see is invoked indirectly by forge_api.
Found only after committing, because .github/scripts/shellcheck-all.sh
derives its lint set from `git ls-files` — an UNTRACKED file is not linted
at all. "Gates clean" measured before `git add` was measuring a set that
excluded the file just written. Verified from a clean clone at the pushed
SHA, which is what caught it.
Refs #188
Term 1's foundation. lib/forge.sh gains forge_select, which sources exactly
one of lib/forge-github.sh or lib/forge-forgejo.sh; both define the same
verbs, so no branching reaches the 61 call sites. The github backend is the
current gh invocation extracted 1:1 — term 5 is kept by making that path
boring.
The page size moves OUT of the call sites and into the backend, because it
is not portable and fails silently. Measured 2026-08-02:
?per_page=100 GitHub 100 items Forgejo 30 items (ignored)
?limit=100 GitHub 30 items Forgejo 50 items (capped)
Both answer HTTP 200 with valid JSON. Every call site here is GitHub-shaped,
so a verbatim port would have swept 30 of rig's 137 issues and printed
"reconciled." — criterion 2 failing green, the same failure class as the
blind sweep. Both page_url helpers strip a stray page-size parameter in
either dialect, so a call site cannot reintroduce it by accident.
Forgejo caps a page at 50 whatever is asked, so pagination is mandatory, not
an optimisation. The gather is then PROVEN complete against x-total-count
rather than assumed complete because a loop ended.
@kimi-reviewer-andresmgsl's hardening (#4699): a missing x-total-count is
itself a loud refusal. Header exposure is a server setting, and an assert
that cannot run must not silently pass — that is the failure class
re-entering through the guard built to stop it.
Call sites are not ported yet; that is the next commit.
Refs #188
@codex-reviewer-andresmgsl's draft-stage finding: the closed/merged half of
the term-3 replacement had an executable-path case, the open half did not.
The 27 closes_references cases test the parser, not the
`.body | @base64` -> base64 -d -> closes_references wiring around it.
Both directions in one sweep so neither assertion passes vacuously: #50 is
closed by an open PR and keeps its claim, #51 is closed by nothing and is
reclaimed. `Closes #50` sits on the third line of the body, so the newline
protection is non-vacuous — an @tsv-shaped regression that keeps only the
first line reclaims #50 and reds the case.
Verified by mutation: replacing the decode with `base64 -d | head -1` fails
exactly "a claim closed by an open PR survives the base64 round trip" and
nothing else; reverting restores 148/148.
The clock is injected. INOW is a fixed 2033 epoch, so without ISSUEFLOW_NOW
the subprocess reads its own wall clock, dates both claims in the future and
keeps them on a negative age — green, and proving nothing. Caught while
writing this case.
Also renames the sibling assertion that still said "through GraphQL"; that
gather has been REST since 5797b41.
Refs #188
Term 3 of #188. Forgejo has no GraphQL API, so these two gathers could not
be translated — there is no endpoint to translate them to. A real
forgejo-runner job says so from the other side: GITHUB_GRAPHQL_URL arrives
set to the empty string (probe task 278).
MERGED_REF_PR_RECORDS was already a body parse; GraphQL was buying
pagination, nothing semantic. OPEN_PR_ISSUES used GitHub's own parse of the
closing keywords, so it becomes lib/closes_references.sh — a sibling of
refs_references, sharing its LOCAL/CROSS classifier so rig#112 can still
never be read as local #112 (#61).
Both gathers now read /pulls, which /api/v3 and /api/v1 return in the same
shape (measured on both). merged_at replaces GraphQL's states: MERGED.
Bodies travel base64: jq's @tsv escapes a newline to a literal backslash-n,
which a line parser reads as one line and loses every declaration after the
first.
The accepted delta, written down rather than rediscovered: GitHub also
records closing links attached through the PR development sidebar, which
live in no body. This family declares links in the body, so the delta is
zero here.
Refs #188
The preflight half of #188, landed first so it stands alone: the forge is
decided once, before any sweep, and a client that cannot speak it exits
non-zero with a named reason.
Measured against forgejo.heavyduty.builders at 84bb1a4 — two of the three
actions reported SUCCESS having read nothing:
labels-scope exit 0 "no .github/labeler.yml" (the file is HTTP 200)
labels-reconcile exit 0 "reconciled." (zero PRs enumerated)
issueflow-reconcile exit 1 "unexpected end of JSON input"
labels-reconcile's blind-sweep warning (#96) could not fire: it counts
unreadable PRs against a list `gh pr list` never produced, and a process
substitution's failure does not trip set -e, so total stayed 0. Installing
gh makes it worse, silencing the one loud failure.
Detection is measured, not inferred from docs: a real forgejo-runner v6.3.1
job (probe task 278) shows Forgejo populating the whole GITHUB_* namespace,
so GITHUB_ACTIONS proves nothing. GITHUB_API_URL's shape, GITEA_ACTIONS and
GITHUB_SERVER_URL do. The same probe shows the runner image carries neither
gh nor stoke, which is what makes the forgejo backend REST.
Tests declare CEREMONY_FORGE at the forge boundary rather than stubbing gh
and staying silent about the forge — the boundary move term 5 asks for.
Refs #188
The reorder let a draft with a live human request plus a standing block
or comment fall through to round_state, whose human-request precedence
sits above BLOCK/FEEDBACK — 224 of claude's 1500 fixture cases read
needs-human on a PR GitHub cannot merge. decide_state now disqualifies
needs-human unconditionally under DRAFT=true, landing on
state:addressing like the blocker/needs-ruling/blocked clauses. The two
new rows assert the criterion where it can actually fail: human
requested x {CHANGES_REQUESTED, COMMENTED}. Also grok's nit: the
bootstrap row for state:building now matches LABELS.md (draft is
evidence, not the definition), and the CONSUMERS.md reflow nits are in.
Refs #205
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
panel[z]]=b parsed at the round-1 head: the case pattern only proves
some ]= occurs, so the stray ] stayed inside the login and the real
author silently fell back to the base panel — the misroute D4 exists to
refuse. The login charset is now enforced with the bracket-specific
diagnostic; codex's probe and an invalid-character row are the new
must-fail fixtures.
Refs #224
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The ruling from discussion #171 as ruled: canonical text in REVIEWER.md
§What you review against item 2 (beside the verify-at-pin sub-bullet it
is the sibling of), short form in BUILDER.md §Building pointing at it.
CONTRIBUTING.md and docs/CONSUMERS.md checked for contradiction or
duplication: none — their pin prose is the mirror/caller pinning rule —
so both are deliberately untouched.
Refs #216
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Each marker now says available-at-tag in the guide's existing L420
phrasing, verified by tag containment in #221; every never-mix-refs
sentence survives verbatim. The convention paragraph gains its missing
half: the release PR that ships machinery clears, in that same PR,
every marker its assembled section makes false.
Refs #221
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
round_outranks_draft consults the round before draft short-circuits: a
re-drafted PR carrying CHANGES_REQUESTED, an owed round-reply, or
push-staled approvals reads state:addressing; a live panel request on a
draft surfaces as state:bots-reviewing rather than being absorbed
(the must-not-paper-over combination, decided as: visible). Approvals do
not outrank draft, so a draft never reads needs-human, and a virgin
draft is byte-identical to before. LABELS.md's state:building row makes
draft evidence, not the definition.
Refs #205
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
One resolution point (panel_for_author) feeds set_required_bots; the
author's row when the conf defines one, the base panel= otherwise, minus
the author in either case. Bracket prefixes are matched quoted so the
case patterns cannot glob (D7, panela= tripwire). configured_label_rows
skips the rows so a dispatch bootstrap cannot mint a label named after
one. BUILDER.md/REVIEWER.md carry the one D9 wording; CONSUMERS.md
publishes the row as unreleased with the parse-failure warning.
Refs #224
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Consume the two displacement-fix fragments into the 0.4.1 section, stamp
VERSION and every CEREMONY_SELF_REF carrier, and record the doors-unchanged
drill ruling with the candidate-head evidence table.
Refs #212
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
crew#250 verified two facts the design prose now carries: a queue-
displaced run is not independently rerunnable (gh run rerun / --failed /
--job all refuse), so a victim PR had no manual escape hatch; and the
displacing burst is deterministic — one review_requested event per
panelist per request — so displacement is the steady state of a working
fleet, scaling with panel size, not a traffic spike.
CONSUMERS.md now walks the adoption as one atomic four-edit PR with crew
as the worked example: the pin bump in every ceremony uses: reference,
the new labels-sweep.yml caller, the cron RELOCATED (bold warning: a
copied-not-moved schedule double-fires sweeps into the one shared group
and reads as the bug getting worse after the fix), and actions: write
replacing the labels caller's actions: read.
Refs #209
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The sweep rode the same workflow run as the PR event that woke it, so
every displacement in the shared labels-reconcile queue recorded a
CANCELLED reconcile check on some PR — fake red CI that held review
requests. The reconcile + issueflow jobs move, unchanged, to a new
reusable labels-sweep.yml behind their own caller; labels.yml gains a
trigger job that dispatches the consumer's sweep caller with the plain
GITHUB_TOKEN (workflow_dispatch is a documented no-retrigger exemption)
on every event that used to run reconcile. A displaced sweep now cancels
on the Actions tab, attached to no PR; PR checks show scope + trigger.
Because every trigger-driven wake arrives as workflow_dispatch, the event
name alone no longer separates the operator's manual bootstrap from an
event-woken sweep: the sweep caller's bootstrap dispatch input does — the
trigger passes no, a bare manual dispatch defaults to yes. The sweep
reusable also takes pr_workflow_name, exported as SELF_WORKFLOW for the
#208 reconciler (harmless to earlier ones; zero file overlap with #208).
The trigger is deliberately loud: a pin bumped without the sweep caller,
its bootstrap input, or actions: write on the labels caller goes red at
the trigger job instead of silently never sweeping again — documented in
docs/CONSUMERS.md with the split stubs and the atomic-adoption note.
Refs #209
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The shared reconcile concurrency group displaces queued sweeps as
CANCELLED, and the displaced run's successor attaches to a different PR —
so on the victim the newest self entry stayed CANCELLED, scored FAILURE,
and the sweep set blocker:ci-red off its own corpse every cadence
(crew#227). Drop rollup entries whose workflowName matches SELF_WORKFLOW
(defaulting to the ambient GITHUB_WORKFLOW — the caller's name, so no
workflow edit and no hardcoded consumer name) before the newest-per-context
collapse; an empty name filters nothing. A self-only rollup now honestly
scores NONE, and a genuine foreign failure still blocks beside a cancelled
self entry — the must-fail guard against re-opening #136.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The trigger comments, CONSUMERS stub, changelog, and reusable labels.yml
comment said "events carry every real state change in seconds; the cron
only backstops a forgotten handoff." That is backwards: no subscribed
event wakes this sweep for a review verdict landing (no pull_request_review
trigger), blocker:ci-red set/cleared, blocker:conflict when another PR
merges, or the time-based stale/48h-reclaim. The hourly cron is the sole
discovery path for those four classes, not a forgotten-handoff net —
so the comments now name them and warn against deleting the cron (AC5).
codex-2 / kimi-2 (both blocking, round 2 @0a812c4b).
Round fixes on #200.
codex-1 (blocking): the issues narrowing dropped `edited`/`reopened`, but both
carry a queue-state change an event uniquely carries — `edited` a body rewrite
of the `Blocked by #N` declaration the sweep parses
(issueflow-reconcile.sh:179), `reopened` a closed issue re-entering the queue.
Dropping them tripped #199's must-fail. Narrow to
`[opened, closed, edited, reopened]`, dropping only the churn/validation
actions labeled/unlabeled/assigned/unassigned. Trigger tests now pin
edited/reopened present and the four dropped; labels.test.sh exact-list updated.
kimi (blocking): the "supersedes unreleased #144" prose was false — #144's
edited/reopened shipped in 0.3.0. Dissolved: we now keep them. CONSUMERS prose
rewritten to the real version history (0.2.0 #32 / 0.3.0 #144 / #199 narrows),
and the #137 review-request line corrected from "unreleased" to shipped-in-0.3.0.
kimi (non-blocking): reusable labels.yml comment no longer cites */15.
codex-2 (AC1 after-measurement / closing) escalated to triage on #199 — held,
not guessed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The tree is grouped (fragments carry ### headings) and changelog-armed
caps each '- ' entry at 300 chars; the single 574-char bullet failed
both. Rewrite as three '### Changed' bullets.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add test/labels-triggers.test.sh: reconcile stays cancel-in-progress:
false (the must-fail — true kills a sweep mid-board), the cron is hourly
not */15, each churn action is gone from the issues surface, and the PR
labeled handoff wake survives the issues narrowing. Update the #137/#144
parity block in labels.test.sh to the narrowed [opened, closed] contract
and replace its fragile inline pull_request_target scan (the #199 prose
comments name the trigger keys) with the anchored event_types reader.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Bring the published caller stub to the #199 shape in lockstep with
self-labels.yml (preserving the stub<->caller parity #144 established):
hourly cron, issues: [opened, closed], per-trigger why-comments. Rewrite
the adoption prose — 0.2.0 shipped the broad set; #199 narrows it at its
pin bump and supersedes the unreleased edited/reopened additions (#144).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The advisory */15 cron fired a full-board sweep four times an hour at
GitHub's 1-minute billing floor; events carry real state changes in
seconds, so the cron only backstops a forgotten handoff. Relax it to
hourly and narrow the issues: trigger to [opened, closed] — the two
actions with promptness-critical reconcile behavior — dropping the
labeled/unlabeled churn source. Per-trigger why-comments added.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Refs #189. Round 1 on #191 (codex blocking, kimi nit).
codex caught that advancing the stamp to crew@4da17c4 made the attention
section false. It did, and the error is mine in kind, not only in detail: a
reconciliation stamp is a claim about the WHOLE file against that tree, and I
audited only the ci-red surface across a 17-commit advance. That is the #187
failure mode, on the PR that closes#189.
What was false at the pinned SHA, all from the crew#66 ruling (danmt,
2026-07-27) landing in d578150e:
- "One wake is registry-independent, by design: attention" — no wake is
exempt now. _attention_partition splits rows against the registry; OUT rows
are reported and never acted on.
- "the assignment is what carries the authorization — there is nothing here
for a repo list to scope" — this is the position the ruling REJECTED. The
cost was argued first: a cross-repo handoff now waits on an operator adding
the repo, which is why an out-of-scope demand also pings the operator over
the boot-gate channel rather than only reaching duty.log.
- "a fix that bounds this wake to the registry would re-create the #16
incident" — that fix landed, with the ping as its mitigation.
- The parenthetical calling crew's repos-default.txt header a contradiction
"raised there as a discussion" — the discussion is crew#66, it was ruled,
and crew's header now names the attention wake explicitly. This file was
preserving the losing side of a settled question.
Also in range and owned by this section: d849f166 ledgered the wake, so
"a session that dies before acking is relaunched; that is the whole
crash-recovery story" no longer is. Dying relaunches; COMPLETING without
acking is a decline, and the ledger stops it re-firing until the issue
moves. Meaning, not mechanism, so it belongs here.
Re-audited the rest of the advance rather than spot-fixing: fleet.roster
still declares itself the TARGET environment (the roster paragraph holds),
and the roster/install commits touch role resolution inside crew, which this
file does not describe.
kimi's nit, which the earlier rewrite made mine: the paper inventory counted
one row while the triage-signals bullet marks a second. Both are needs-ruling
rows and both are now named — a number that has to be recounted every time a
wake lands is the thing that went wrong.
18/18 test files; changelog_fragment_problem OK; shellcheck, self-ref and
git diff --check clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Refs #189 — the post-merge, triage-owned task, now reachable: heavy-duty/crew#64
merged at 4da17c49594c2d86bd3793fa3567846cbca38e90.
#190 wrote the ci-red wake into FLEET.md deliberately marked on paper, per this
file's convention for a spec that is not yet running, and left the reconciliation
stamp at crew@01fb49c because advancing it to a SHA that did not carry the wake
would be the drift #187 exists to remove. crew#64 has merged, so both halves flip:
- The stamp advances to crew@4da17c4, and every crew permalink in the file with
it. A stamp and its evidence links naming different trees is worse than no
stamp: the reader diffs the wrong engine and finds no drift because the tree
they were pointed at is the one the prose was written against.
- The four on-paper markers go: the duty-order caveat, the ci-red bullet's
parenthetical, the Build bullet's "once ci-red deploys", and the closing
paper inventory — which now names one remaining paper wake, the notifier's
needs-ruling queue, not two.
The Build bullet is not a pure marker removal. crew#64's last review round
changed what it has to say: the operator ruled the round gate a whitelist, so
the wake admits a green head OR one with no checks configured, and holds a red
head AND one whose check has not finished. Copying the old "never a round at a
red head" through the flip would have shipped a fresh inaccuracy on the same
commit that claims the file is reconciled. The ci-red bullet gains the matching
sentence from the other side: an unfinished check is not a red head and wakes
nothing, because nothing has failed yet.
Verified at the stamped SHA rather than assumed: duty.sh's header carries
attention → … → resume → ci-red → build (and says it is what this file is
reconciled against), shared/README.md's duty order matches, and notify.sh's
only label filter is still state:needs-human — which is what keeps the
remaining paper claim true.
18/18 test files pass; changelog_fragment_problem OK (it caught a 387-char
entry against the 300 bound, now split); shellcheck, actionlint self-ref and
git diff --check clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The round protocol told the author to re-request exactly the reviewers
who did not approve, but the handoff predicate counts an approval only
at the current head: any push while answering a round staled the
earlier approver's verdict, doctrine said not to re-request them, and
the PR could never converge — the silent-stall shape of #26/#39.
Step 2 now re-requests by head, not by verdict: every panelist after a
push, the non-approvers alone at an unchanged head. Shape 2's wording
is aligned so the two paragraphs agree.
Defect raised by dan-claude-bot on #190; folded in at the operator's
direction while the paragraph is open. Refs #189.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
