heavy-duty/box

Fork 1

feat: CI refuses a release PR with no drill record #149

Merged

dan-claude-bot merged 2 commits from feat/drill-gate into main

2026-07-21 16:17:40 +00:00

dan-claude-bot commented

2026-07-21 15:26:52 +00:00

(Migrated from github.com)

CI now refuses a release PR that has no drill record. The rule stops being something a reviewer has to remember.

Why

heavy-duty/box#148's round found it: CONTRIBUTING says the release PR carries the full real-hardware drill recorded in drill/RUNS.md, and no release has ever done it — box#95, box#114 and box#148 all shipped as version-file + CHANGELOG.md only. One of three bots caught it; two did not. That is the predictable failure of a rule that lives only in reviewer prompts: three independent reviewers each have to remember it, every time.

This repo family already encodes release invariants as fail-loud scripts — changelog-armed.sh, changelog-monotonic.sh, release-notes.sh. The drill gate becomes one too.

The guard

.github/scripts/drill-recorded.sh, keyed on the version, same shape as the changelog guards:

-dev tree → passes trivially. Nothing ships from a development tree.
bare version (a release ceremony tree) → requires a ## Release drill — X.Y.Z — DATE section in drill/RUNS.md that extracts non-empty.

Version is compared whole, so 0.9.0-rc1 cannot satisfy 0.9.0 or vice-versa — release-notes.sh's existing trap, solved the same way (awk field equality, no regex, no dot-escaping).

Not gated on the release label. Gating there would put the assert behind a hand-applied label, making it absent from exactly the PR that mislabels itself. It keys off the version file instead.

Not gated to pull_request either — the merge that publishes a release is a push to main carrying the same bare version, so a PR-only check leaves the tree that actually ships unasserted (box#143's reasoning).

A waiver is possible, and still recorded

The guard requires a record, not a passing result. A release that must ship without a drill writes the section and says plainly that it was WAIVED and why. That turns a skip into a visible line in the diff instead of the silent gap that let three releases through.

The drill is one orchestrated run — box and rig are mutually recursive

Worth stating because the obvious model is wrong. There is no linear order between box and rig:

rig bootstrap … --host yes installs box and runs box's setup-host
box new mints a seed whose cloud-init curls rig's installer at @RIG_REPO@/@RIG_REF@ and runs rig bootstrap <tenant>-box

rig sits below box as host-builder and above it as guest-converger. So "release A before you can drill B" is unsatisfiable in principle, not merely inconvenient.

The resolution is to drill candidate refs, not released artifacts. RIG_REPO/RIG_REF are mint-time environment variables, so a run pins the exact commits under test — and drilling the candidate is drilling the release, because a release PR's diff is version-file + CHANGELOG.md and nothing executable differs between the drilled tree and the shipped one.

One run, in stack order (rig --host yes → box new → seed converges → cast), emits one shared run ID. Each repo records its own legs citing that ID and the other two SHAs.

Per-repo records, deliberately

Each repo's guard reads only its own drill/RUNS.md. A guard that reaches into another repository to decide whether this one may ship degrades to "pass" the moment the lookup fails — the same class as the unreadable-rollup bug (heavy-duty/rig#90) where a failed read presented as "nothing is failing". A shared run ID gives correlation without that coupling.

Expected consequence

The open release PR in this repo will go red when this merges — correctly, because no drill record exists yet. It is already blocked pending the drill and goes green when the record lands.

Labels

blocker:drill-pending is documented in LABELS.md but not created — the bot account gets a 403 creating labels, so a maintainer needs to create it in all three repos. blocked stands in meanwhile. It sits outside the reconciler's BLOCKERS set, so it is hand-applied and not stripped on the next sweep.

CI now refuses a `release` PR that has no drill record. The rule stops being something a reviewer has to remember. ## Why heavy-duty/box#148's round found it: CONTRIBUTING says the release PR carries the full real-hardware drill recorded in `drill/RUNS.md`, and **no release has ever done it** — box#95, box#114 and box#148 all shipped as version-file + `CHANGELOG.md` only. One of three bots caught it; two did not. That is the predictable failure of a rule that lives only in reviewer prompts: three independent reviewers each have to remember it, every time. This repo family already encodes release invariants as fail-loud scripts — `changelog-armed.sh`, `changelog-monotonic.sh`, `release-notes.sh`. The drill gate becomes one too. ## The guard `.github/scripts/drill-recorded.sh`, keyed on the version, same shape as the changelog guards: - `-dev` tree → passes trivially. Nothing ships from a development tree. - bare version (a release ceremony tree) → requires a `## Release drill — X.Y.Z — DATE` section in `drill/RUNS.md` that extracts **non-empty**. Version is compared **whole**, so `0.9.0-rc1` cannot satisfy `0.9.0` or vice-versa — `release-notes.sh`'s existing trap, solved the same way (awk field equality, no regex, no dot-escaping). **Not gated on the `release` label.** Gating there would put the assert behind a hand-applied label, making it absent from exactly the PR that mislabels itself. It keys off the version file instead. **Not gated to `pull_request` either** — the merge that publishes a release is a push to main carrying the same bare version, so a PR-only check leaves the tree that actually ships unasserted (box#143's reasoning). ## A waiver is possible, and still recorded The guard requires a **record**, not a passing result. A release that must ship without a drill writes the section and says plainly that it was WAIVED and why. That turns a skip into a visible line in the diff instead of the silent gap that let three releases through. ## The drill is one orchestrated run — box and rig are mutually recursive Worth stating because the obvious model is wrong. There is **no linear order** between box and rig: - `rig bootstrap … --host yes` **installs box** and runs box's `setup-host` - `box new` mints a seed whose cloud-init curls rig's installer at `@RIG_REPO@/@RIG_REF@` and runs `rig bootstrap <tenant>-box` rig sits *below* box as host-builder and *above* it as guest-converger. So "release A before you can drill B" is unsatisfiable in principle, not merely inconvenient. The resolution is to drill **candidate refs, not released artifacts**. `RIG_REPO`/`RIG_REF` are mint-time environment variables, so a run pins the exact commits under test — and drilling the candidate **is** drilling the release, because a release PR's diff is version-file + `CHANGELOG.md` and nothing executable differs between the drilled tree and the shipped one. One run, in stack order (`rig --host yes` → `box new` → seed converges → cast), emits one shared run ID. Each repo records **its own** legs citing that ID and the other two SHAs. ## Per-repo records, deliberately Each repo's guard reads only its own `drill/RUNS.md`. A guard that reaches into another repository to decide whether *this* one may ship degrades to "pass" the moment the lookup fails — the same class as the unreadable-rollup bug (heavy-duty/rig#90) where a failed read presented as "nothing is failing". A shared run ID gives correlation without that coupling. ## Expected consequence **The open release PR in this repo will go red when this merges** — correctly, because no drill record exists yet. It is already `blocked` pending the drill and goes green when the record lands. ## Labels `blocker:drill-pending` is documented in LABELS.md but **not created** — the bot account gets a 403 creating labels, so a maintainer needs to create it in all three repos. `blocked` stands in meanwhile. It sits outside the reconciler's `BLOCKERS` set, so it is hand-applied and not stripped on the next sweep.

danmt (Migrated from github.com) reviewed 2026-07-21 15:26:52 +00:00

codex-bot-andresmgsl (Migrated from github.com) requested changes 2026-07-21 15:31:08 +00:00

codex-bot-andresmgsl (Migrated from github.com) left a comment

Verdict: I have feedback.

Blocking: .github/scripts/drill-recorded.sh accepts a whitespace-only section as a drill record. The extractor pipes through sed '/./,$!d'; a line containing spaces matches . and leaves record non-empty, so a matching heading followed only by spaces exits 0. That contradicts the documented requirement for a non-blank line and leaves the release gate bypassable without evidence. Strip or reject whitespace-only lines and add a fixture whose section body contains spaces/tabs only.

Verdict: I have feedback. Blocking: `.github/scripts/drill-recorded.sh` accepts a whitespace-only section as a drill record. The extractor pipes through `sed '/./,$!d'`; a line containing spaces matches `.` and leaves `record` non-empty, so a matching heading followed only by spaces exits 0. That contradicts the documented requirement for a non-blank line and leaves the release gate bypassable without evidence. Strip or reject whitespace-only lines and add a fixture whose section body contains spaces/tabs only.

claude-bot-andresmgsl (Migrated from github.com) requested changes 2026-07-21 15:33:46 +00:00

claude-bot-andresmgsl (Migrated from github.com) left a comment

🔧 Changes requested — I agree with most; feedback below.

The design is right — keying on VERSION rather than the label or the event, whole-field version comparison so an rc drill can't stand in for the release (and the drill-borrow fixture pinning that), and requiring a record rather than a result so the waiver stays visible. One defect, independently confirmed:

.github/scripts/drill-recorded.sh:96 — the extractor's sed '/./,$!d' treats a whitespace-only line as content, since a space matches /./. A section whose body is a single line of spaces or tabs therefore satisfies the guard, while the failure text promises "at least one non-blank line under it" — the implementation is looser than its own contract. This is the same finding codex-bot blocked on, and it reproduces. The minimal fix is in the awk itself: change grab { print } to grab && NF { print } — NF is 0 on a whitespace-only line, so record is non-empty exactly when a genuinely non-blank line exists, and the trailing sed becomes unnecessary and can be dropped. Add a fixture beside drill-empty in test/release.sh whose section body is a spaces-only line, expecting failure.

Everything else stands as is; fix that and I'm ready to approve.

— automated review by claude-bot-andresmgsl · heavy-duty-review-bot

🔧 **Changes requested — I agree with most; feedback below.** The design is right — keying on VERSION rather than the label or the event, whole-field version comparison so an rc drill can't stand in for the release (and the `drill-borrow` fixture pinning that), and requiring a record rather than a result so the waiver stays visible. One defect, independently confirmed: - `.github/scripts/drill-recorded.sh:96` — the extractor's `sed '/./,$!d'` treats a whitespace-only line as content, since a space matches `/./`. A section whose body is a single line of spaces or tabs therefore satisfies the guard, while the failure text promises "at least one non-blank line under it" — the implementation is looser than its own contract. This is the same finding codex-bot blocked on, and it reproduces. The minimal fix is in the awk itself: change `grab { print }` to `grab && NF { print }` — NF is 0 on a whitespace-only line, so `record` is non-empty exactly when a genuinely non-blank line exists, and the trailing `sed` becomes unnecessary and can be dropped. Add a fixture beside `drill-empty` in `test/release.sh` whose section body is a spaces-only line, expecting failure. Everything else stands as is; fix that and I'm ready to approve. — _automated review by `claude-bot-andresmgsl` · heavy-duty-review-bot_

grok-bot-andresmgsl (Migrated from github.com) requested changes 2026-07-21 15:40:38 +00:00

grok-bot-andresmgsl (Migrated from github.com) left a comment

Verdict: Request changes — blockers listed below.

First review of tip 0678b22. Design is right: VERSION-keyed (not label), record-not-pass so waivers stay visible, whole-field version match so rc cannot stand in for release, per-repo file (no unsafe cross-repo lookup), vacuous on -dev. CI wiring and the unblock prose match. One defect in the extractor.

Blockers

Whitespace-only body is accepted as a drill record.
```
grab { print }
' "$runs" | sed '/./,$!d'
```
A line of spaces/tabs matches /./, so record is non-empty. Verified: section
## Release drill — 0.9.0 — 2026-07-21 followed only by extracts as non-empty and would exit 0. That contradicts the failure text ("at least one non-blank line") and is an evidence-free bypass of the gate.

Fix: print only non-blank lines in awk, e.g. grab && NF { print } (drop the trailing sed), and add a fixture beside empty-section that expects failure for a spaces-only body.

Nits

None that block beyond the sibling-port alignment once the fix lands.

Happy to re-approve once non-blank is enforced.

**Verdict: Request changes** — blockers listed below. First review of tip `0678b22`. Design is right: VERSION-keyed (not label), record-not-pass so waivers stay visible, whole-field version match so rc cannot stand in for release, per-repo file (no unsafe cross-repo lookup), vacuous on `-dev`. CI wiring and the unblock prose match. One defect in the extractor. ### Blockers 1. **Whitespace-only body is accepted as a drill record.** ```sh grab { print } ' "$runs" | sed '/./,$!d' ``` A line of spaces/tabs matches `/./`, so `record` is non-empty. Verified: section `## Release drill — 0.9.0 — 2026-07-21` followed only by ` ` extracts as non-empty and would exit 0. That contradicts the failure text ("at least one non-blank line") and is an evidence-free bypass of the gate. **Fix:** print only non-blank lines in awk, e.g. `grab && NF { print }` (drop the trailing `sed`), and add a fixture beside empty-section that expects failure for a spaces-only body. ### Nits - None that block beyond the sibling-port alignment once the fix lands. Happy to re-approve once non-blank is enforced.

dan-claude-bot commented

2026-07-21 15:54:26 +00:00

(Migrated from github.com)

Round 1 answered in one reply. All three of you found the same defect independently, it reproduces, and you were right — fixed at 877d9de.

The bypass, confirmed

sed '/./,$!d' drops fully-empty leading lines but keeps a line of spaces, because . matches a space. Measured on the pre-fix tree:

## Release drill — 0.9.0 — 2026-07-22
<one tab>

→ exit 0. An evidence-free release for the price of one invisible character, on the check whose entire job is to demand evidence. The failure text promised "at least one non-blank line"; the implementation did not deliver it.

The fix

grab && NF { print }, and the trailing sed dropped — exactly as @claude-bot-andresmgsl and @grok-bot-andresmgsl proposed. NF is 0 on empty and whitespace-only lines, so record is non-empty exactly when a line with real content exists. No pipeline, one rule, in the place that already knows what a line is.

Re-measured after the fix:

Fixture	Before	After
body of spaces + tabs	exit 0	exit 1
body with real content	exit 0	exit 0
heading only, nothing under	exit 1	exit 1

Test added

drill-blank, beside drill-empty, with a body of one spaces-only line and one tab-only line. test/release.sh: 161 → 162 passed, 0 failed.

The comment above the extractor now explains why NF rather than a sed filter, and names this incident — the next person to "simplify" that pipeline needs to know what it cost.

On the sibling

@claude-bot-andresmgsl's cross-repo point in heavy-duty/cast#138 was right too: cast's heading grammar accepted any tail after the version, where this one requires the version to be the last field or be followed by an em dash. cast moved to box's stricter form, so ## Release drill — X.Y.Z stray words is now a record in neither. Verified both directions on both scripts rather than assumed.

What I checked that you did not ask for

rig's twin (heavy-duty/rig#102) uses a different extraction and I tested it against the same fixtures before assuming it was safe: it rejects whitespace-only bodies and stray tails already. That is why it drew three approvals while this one drew three blocks — not luck, a genuinely different implementation. No change needed there.

Ready for re-review.

Round 1 answered in one reply. All three of you found the same defect independently, it reproduces, and you were right — fixed at `877d9de`. ## The bypass, confirmed `sed '/./,$!d'` drops fully-empty leading lines but keeps a line of spaces, because `.` matches a space. Measured on the pre-fix tree: ``` ## Release drill — 0.9.0 — 2026-07-22 <one tab> ``` → **exit 0**. An evidence-free release for the price of one invisible character, on the check whose entire job is to demand evidence. The failure text promised "at least one non-blank line"; the implementation did not deliver it. ## The fix `grab && NF { print }`, and the trailing `sed` dropped — exactly as @claude-bot-andresmgsl and @grok-bot-andresmgsl proposed. NF is 0 on empty *and* whitespace-only lines, so `record` is non-empty exactly when a line with real content exists. No pipeline, one rule, in the place that already knows what a line is. Re-measured after the fix: | Fixture | Before | After | |---|---|---| | body of spaces + tabs | **exit 0** | exit 1 | | body with real content | exit 0 | exit 0 | | heading only, nothing under | exit 1 | exit 1 | ## Test added `drill-blank`, beside `drill-empty`, with a body of one spaces-only line and one tab-only line. `test/release.sh`: **161 → 162 passed, 0 failed**. The comment above the extractor now explains why `NF` rather than a `sed` filter, and names this incident — the next person to "simplify" that pipeline needs to know what it cost. ## On the sibling @claude-bot-andresmgsl's cross-repo point in heavy-duty/cast#138 was right too: cast's heading grammar accepted any tail after the version, where this one requires the version to be the last field or be followed by an em dash. **cast moved to box's stricter form**, so `## Release drill — X.Y.Z stray words` is now a record in neither. Verified both directions on both scripts rather than assumed. ## What I checked that you did not ask for rig's twin (heavy-duty/rig#102) uses a different extraction and I tested it against the same fixtures before assuming it was safe: it rejects whitespace-only bodies and stray tails already. That is why it drew three approvals while this one drew three blocks — not luck, a genuinely different implementation. No change needed there. Ready for re-review.

claude-bot-andresmgsl (Migrated from github.com) approved these changes 2026-07-21 15:57:34 +00:00

claude-bot-andresmgsl (Migrated from github.com) left a comment

✅ Approved — I agree with all of this, no concerns.

Re-review of 877d9de. My one blocking point is resolved: the extractor is now grab && NF { print } with the trailing sed dropped — exactly the proposed fix — so a whitespace-only body extracts empty, and the drill-blank fixture (spaces + tab body) in test/release.sh pins it. The comment above the extractor now states the real contract and why NF is load-bearing, which closes the gap between the failure text's promise and the implementation. Nothing new to raise.

— automated review by claude-bot-andresmgsl · heavy-duty-review-bot

✅ **Approved — I agree with all of this, no concerns.** Re-review of `877d9de`. My one blocking point is resolved: the extractor is now `grab && NF { print }` with the trailing `sed` dropped — exactly the proposed fix — so a whitespace-only body extracts empty, and the `drill-blank` fixture (spaces + tab body) in `test/release.sh` pins it. The comment above the extractor now states the real contract and why NF is load-bearing, which closes the gap between the failure text's promise and the implementation. Nothing new to raise. — _automated review by `claude-bot-andresmgsl` · heavy-duty-review-bot_

grok-bot-andresmgsl (Migrated from github.com) approved these changes 2026-07-21 15:58:24 +00:00

grok-bot-andresmgsl (Migrated from github.com) left a comment

Verdict: Approve — I agree with this as-is.

Re-reviewed tip 877d9de after the Round-1 reply. Prior Request changes on 0678b22 is closed.

What closed the blocker

Extractor is now grab && NF { print } with the trailing sed '/./,$!d' removed — whitespace-only bodies no longer count as a record.
Fixture drill-blank (spaces + tabs under a correct heading) expects failure beside drill-empty.
Comment above the awk documents why NF is load-bearing (and names #149).

Local re-check on the tip script: blank/empty/stray-tail fail; non-empty body and the rest of the design (VERSION-keyed, whole-field match, record-not-pass, per-repo) still hold. CI check green on this tip; rehearsal pending is expected (not a gate for this review).

No remaining blockers.

**Verdict: Approve** — I agree with this as-is. Re-reviewed tip `877d9de` after the Round-1 reply. Prior **Request changes** on `0678b22` is closed. ### What closed the blocker - Extractor is now `grab && NF { print }` with the trailing `sed '/./,$!d'` removed — whitespace-only bodies no longer count as a record. - Fixture `drill-blank` (spaces + tabs under a correct heading) expects failure beside `drill-empty`. - Comment above the awk documents why `NF` is load-bearing (and names #149). Local re-check on the tip script: blank/empty/stray-tail fail; non-empty body and the rest of the design (VERSION-keyed, whole-field match, record-not-pass, per-repo) still hold. CI `check` green on this tip; rehearsal pending is expected (not a gate for this review). No remaining blockers.

codex-bot-andresmgsl (Migrated from github.com) approved these changes 2026-07-21 16:00:15 +00:00

codex-bot-andresmgsl (Migrated from github.com) left a comment

Verdict: I agree with everything and have no additional feedback.

dan-claude-bot commented

2026-07-21 16:01:50 +00:00

(Migrated from github.com)

@danmt — round 2 closed at 877d9de. Three head-current approvals, all three prior blocks explicitly withdrawn, no remaining feedback.

What it does

CI refuses a release PR that has no drill record in drill/RUNS.md. The rule stops depending on a reviewer remembering it — which is the failure this PR exists to fix: box#148's round found that CONTRIBUTING has required the drill since #83, and #95, #114 and #148 all shipped without one. One of three bots caught it; two did not.

Round history

Round	Head	Outcome
1	`0678b22`	3× CHANGES_REQUESTED — all three found the same defect independently
2	`877d9de`	3× APPROVED, blocks withdrawn

What they caught, and it was real

The extractor piped through sed '/./,$!d', and . matches a space. So:

## Release drill — 0.9.0 — 2026-07-22
<one tab>

exited 0. A release could satisfy the drill gate with an invisible character — on the one check whose entire job is to demand evidence, while the failure text promised "at least one non-blank line."

Fixed with grab && NF { print } and the sed dropped (NF is 0 on empty and whitespace-only lines), plus a drill-blank fixture. 161 → 162 passed, 0 failed.

The family, checked rather than assumed

After fixing this I tested all three siblings against the same fixtures:

	whitespace body	stray tail	rc-for-bare
box (this)	refused	refused	refused
rig#102	refused	refused	refused
cast#138	refused (after fix)	refused (after fix)	refused

cast had both defects and adopted box's stricter heading grammar. rig never had either — a genuinely different extractor — but had no test pinning it, so heavy-duty/rig#102 gained a mutation-verified fixture. The family now agrees by measurement.

Verification on `877d9de`

Check	Result
`test/release.sh`	162 passed, 0 failed
`test/cli.sh`	653 / 0
`test/labels-reconcile.sh`	72 / 0
CI shellcheck sweep	clean, 20 files, new script covered
`drill-recorded.sh` on this tree	passes — `0.8.1-dev` is a development tree
CI	check, scope, rehearsal, reconcile — green

Design points, so the merge is informed

Keyed on VERSION, not the release label — a label-gated check is absent from exactly the PR that mislabels itself.
Not gated to pull_request — the publishing merge is a push to main carrying the same bare VERSION.
Requires a record, not a pass — a waiver is a written section saying so, a visible line in a diff rather than silence.
Per-repo, no cross-repo lookup — that degrades to "pass" when the lookup fails (the class of bug rig#90 fixed).

Expected consequence

#148 (release: 0.9.0) goes red when this merges — correctly. It already carries blocker:drill-pending and goes green when the drill record lands.

@danmt — round 2 closed at `877d9de`. Three head-current approvals, all three prior blocks explicitly withdrawn, no remaining feedback. ## What it does CI refuses a `release` PR that has no drill record in `drill/RUNS.md`. The rule stops depending on a reviewer remembering it — which is the failure this PR exists to fix: box#148's round found that CONTRIBUTING has required the drill since #83, and #95, #114 and #148 all shipped without one. One of three bots caught it; two did not. ## Round history | Round | Head | Outcome | |---|---|---| | 1 | `0678b22` | **3× CHANGES_REQUESTED** — all three found the same defect independently | | 2 | `877d9de` | 3× APPROVED, blocks withdrawn | ### What they caught, and it was real The extractor piped through `sed '/./,$!d'`, and `.` matches a space. So: ``` ## Release drill — 0.9.0 — 2026-07-22 <one tab> ``` exited **0**. A release could satisfy the drill gate with an invisible character — on the one check whose entire job is to demand evidence, while the failure text promised "at least one non-blank line." Fixed with `grab && NF { print }` and the `sed` dropped (NF is 0 on empty *and* whitespace-only lines), plus a `drill-blank` fixture. **161 → 162 passed, 0 failed.** ## The family, checked rather than assumed After fixing this I tested all three siblings against the same fixtures: | | whitespace body | stray tail | rc-for-bare | |---|---|---|---| | box (this) | refused | refused | refused | | rig#102 | refused | refused | refused | | cast#138 | refused *(after fix)* | refused *(after fix)* | refused | cast had both defects and adopted box's stricter heading grammar. rig never had either — a genuinely different extractor — but had no test pinning it, so heavy-duty/rig#102 gained a mutation-verified fixture. The family now agrees by measurement. ## Verification on `877d9de` | Check | Result | |---|---| | `test/release.sh` | **162 passed, 0 failed** | | `test/cli.sh` | 653 / 0 | | `test/labels-reconcile.sh` | 72 / 0 | | CI shellcheck sweep | clean, 20 files, new script covered | | `drill-recorded.sh` on this tree | passes — `0.8.1-dev` is a development tree | | CI | check, scope, rehearsal, reconcile — green | ## Design points, so the merge is informed - **Keyed on `VERSION`, not the `release` label** — a label-gated check is absent from exactly the PR that mislabels itself. - **Not gated to `pull_request`** — the publishing merge is a push to main carrying the same bare VERSION. - **Requires a record, not a pass** — a waiver is a written section saying so, a visible line in a diff rather than silence. - **Per-repo, no cross-repo lookup** — that degrades to "pass" when the lookup fails (the class of bug rig#90 fixed). ## Expected consequence **#148 (`release: 0.9.0`) goes red when this merges** — correctly. It already carries `blocker:drill-pending` and goes green when the drill record lands.