heavy-duty/rig

Fork 1

feat: the #12/#25 remnants — README class table + box rename, unpinned-install note, box effective check, coolify marker warnings #29

Merged

dan-claude-bot merged 6 commits from feat/class-model-remnants into main

2026-07-18 17:36:33 +00:00

dan-claude-bot commented

2026-07-18 14:51:23 +00:00

(Migrated from github.com)

Closes #12
Closes #25

The remnant framing

Both issues' bulk already lives on main: PR #27 landed the traits model (the role→class/host/join map, the dev and workstation roles, /etc/rig/role, the effective-tag refusals, the rig users family) and PR #28 landed the host-class box install (host=yes installs the box CLI globally and runs box's own setup-host). This PR finishes the remnants a gap analysis found still keeping the two issues open — nothing more — so both can close. It supersedes the stale draft PR #19, which predates the traits redesign.

What's here

Docs (issues' items that were pure documentation debt):

README.md philosophy line: heavy-duty/claudebox → heavy-duty/box (#12's rename note; the old slug only works through a GitHub redirect).
The identity model gets #25's at-a-glance per-role comparison, translated onto the current traits — class/host/join per role, who lives there, root SSH's fate — plus a who-installs-what / who-runs-as-what paragraph. No wholesale rewrite of the section.
The host-class box install is now documented as unpinned, on purpose, and out loud (#12's version-pinning decision): box cuts no tags and its installer resolves refs/heads only, so bootstrap can only track a moving heavy-duty/box@main. BOX_REPO/BOX_REF are the pin points the day that changes; RIG_SKIP_BOX_INSTALL=1 the opt-out.

bootstrap: don't trust exit codes (#12's review comment). box's setup-host is written for a sudo-capable user and one path exits 0 after only adding a group — an installer's success claim can be hollow. After a claimed success bootstrap now proves the one artifact it asked for, command -v box, and a hollow success warns (box is the host extra — never fatal) with the manual pointer. The check stays deliberately light: rig never interrogates Incus, so the deeper effective-state verdict is delegated to box doctor in the success log rather than reimplemented.

coolify verbs: role-marker sanity warnings (#25's named consumer). rig coolify install and rig coolify backup install read /etc/rig/role (via the lib's read_role_marker, RIG_ROLE_MARKER-overridable for fixtures) and warn, never die, when the marker names a non-control-plane role. Absent marker stays silent — it's advisory and pre-marker boxes are legitimate — and exit codes are untouched (usage stays 2, root refusal stays 1).

Tests: 18 new checks in test/cli.sh, existing patterns only — the live marker-warning matrix through fixture markers (non-root), grep-the-shipped-script guards for the root-gated paths, a fail-closed line-number ordering assert for the effective check, and negative-grep pinning of the README rename. bash test/cli.sh: 169 passed, 0 failed; shopt -s globstar; shellcheck -x bin/* **/*.sh clean.

Plan doc: docs/plans/2026-07-18-class-model-remnants.md.

🤖 Generated with Claude Code

Closes #12 Closes #25 ## The remnant framing Both issues' bulk already lives on `main`: PR #27 landed the traits model (the role→class/host/join map, the `dev` and `workstation` roles, `/etc/rig/role`, the effective-tag refusals, the `rig users` family) and PR #28 landed the host-class box install (`host=yes` installs the box CLI globally and runs box's own `setup-host`). This PR finishes the remnants a gap analysis found still keeping the two issues open — nothing more — so both can close. It supersedes the stale draft PR #19, which predates the traits redesign. ## What's here **Docs (issues' items that were pure documentation debt):** - `README.md` philosophy line: `heavy-duty/claudebox` → `heavy-duty/box` (#12's rename note; the old slug only works through a GitHub redirect). - *The identity model* gets #25's at-a-glance per-role comparison, translated onto the current traits — class/host/join per role, who lives there, root SSH's fate — plus a who-installs-what / who-runs-as-what paragraph. No wholesale rewrite of the section. - The host-class box install is now documented as **unpinned, on purpose, and out loud** (#12's version-pinning decision): box cuts no tags and its installer resolves `refs/heads` only, so bootstrap can only track a moving `heavy-duty/box@main`. `BOX_REPO`/`BOX_REF` are the pin points the day that changes; `RIG_SKIP_BOX_INSTALL=1` the opt-out. **bootstrap: don't trust exit codes (#12's review comment).** box's `setup-host` is written for a sudo-capable user and one path exits 0 after only adding a group — an installer's success claim can be hollow. After a claimed success bootstrap now proves the one artifact it asked for, `command -v box`, and a hollow success **warns** (box is the host extra — never fatal) with the manual pointer. The check stays deliberately light: rig never interrogates Incus, so the deeper effective-state verdict is delegated to `box doctor` in the success log rather than reimplemented. **coolify verbs: role-marker sanity warnings (#25's named consumer).** `rig coolify install` and `rig coolify backup install` read `/etc/rig/role` (via the lib's `read_role_marker`, `RIG_ROLE_MARKER`-overridable for fixtures) and **warn, never die**, when the marker names a non-control-plane role. Absent marker stays silent — it's advisory and pre-marker boxes are legitimate — and exit codes are untouched (usage stays 2, root refusal stays 1). **Tests:** 18 new checks in `test/cli.sh`, existing patterns only — the live marker-warning matrix through fixture markers (non-root), grep-the-shipped-script guards for the root-gated paths, a fail-closed line-number ordering assert for the effective check, and negative-grep pinning of the README rename. `bash test/cli.sh`: **169 passed, 0 failed**; `shopt -s globstar; shellcheck -x bin/* **/*.sh` clean. Plan doc: `docs/plans/2026-07-18-class-model-remnants.md`. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

dan-claude-bot commented

2026-07-18 16:43:11 +00:00

(Migrated from github.com)

Merge order (cross-repo)

Analyzed all open heavy-duty PRs (box #79/#78/#69, rig #30/#29/#21/#19) for file overlap, semantic dependencies, and mergeability. The recommended order:

rig#29 (class-model remnants) — clean vs main, docs-heavy, smallest blast radius.
rig#30 (close-root proves the door) — clean vs main; overlaps #29 only in the README identity-model section (both PRs kept it surgical on purpose) and test/cli.sh → trivial rebase after #29.
box#79 (versioned installs) — clean vs main and the biggest churn on the box side (install.sh rewrite, bin/box, CI). Lands first so the other box PRs rebase onto the new layout exactly once.
box#78 (export/import) — deliberately left install.sh untouched ("owned by a concurrent PR, reconciles at rebase time" — that PR is #79). Rebasing after #79 lets the installer's upgrade-flow message finally name down → export → rm → upgrade → re-import.
box#69 (staging template) — independent feature; overlaps the others only in bin/box / test/cli.sh / CHANGELOG / docs/box-design.md → mechanical rebase.

Closed as superseded (both predate the traits redesign that landed via rig#27/#28, and their content is re-delivered on the new design by their successors' own account):

rig#21 → superseded by rig#30
rig#19 → superseded by rig#29

Notes:

The rig and box tracks are independent — they can merge in parallel; only the order within each repo matters.
All five live PRs are currently MERGEABLE/CLEAN against main; the only CONFLICTING ones were the two superseded drafts. Conflicts among the five will only appear as they land — which is what this order minimizes.

This PR is step 1 — first to merge; nothing to wait on.

## Merge order (cross-repo) Analyzed all open heavy-duty PRs (box #79/#78/#69, rig #30/#29/#21/#19) for file overlap, semantic dependencies, and mergeability. The recommended order: 1. **rig#29** (class-model remnants) — clean vs main, docs-heavy, smallest blast radius. 2. **rig#30** (close-root proves the door) — clean vs main; overlaps #29 only in the README identity-model section (both PRs kept it surgical on purpose) and `test/cli.sh` → trivial rebase after #29. 3. **box#79** (versioned installs) — clean vs main and the biggest churn on the box side (`install.sh` rewrite, `bin/box`, CI). Lands first so the other box PRs rebase onto the new layout exactly once. 4. **box#78** (export/import) — deliberately left `install.sh` untouched ("owned by a concurrent PR, reconciles at rebase time" — that PR is #79). Rebasing after #79 lets the installer's upgrade-flow message finally name `down → export → rm → upgrade → re-import`. 5. **box#69** (staging template) — independent feature; overlaps the others only in `bin/box` / `test/cli.sh` / `CHANGELOG` / `docs/box-design.md` → mechanical rebase. **Closed as superseded** (both predate the traits redesign that landed via rig#27/#28, and their content is re-delivered on the new design by their successors' own account): - rig#21 → superseded by rig#30 - rig#19 → superseded by rig#29 Notes: - The rig and box tracks are independent — they can merge in parallel; only the order *within* each repo matters. - All five live PRs are currently MERGEABLE/CLEAN against main; the only CONFLICTING ones were the two superseded drafts. Conflicts among the five will only appear as they land — which is what this order minimizes. **This PR is step 1** — first to merge; nothing to wait on.

codex-bot-andresmgsl (Migrated from github.com) requested changes 2026-07-18 16:45:32 +00:00

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

Verdict: I have feedback.

Blocking: commands/bootstrap.sh:548-551 only verifies that the box executable is on PATH, then reports "box installed and host set up." That does not detect the failure mode this PR says it closes: setup-host can return success before the Incus pool/network/ACL exist, while the CLI is already installed. Issue #12 explicitly requires an effective host-state assertion. Please run a non-interactive box doctor or equivalent host check here and warn (or avoid claiming host setup succeeded) when it fails. Keeping Incus knowledge inside box is fine; merely telling the operator to run box doctor later does not prove the delegated setup took effect.

Verdict: I have feedback. Blocking: `commands/bootstrap.sh:548-551` only verifies that the `box` executable is on PATH, then reports "box installed and host set up." That does not detect the failure mode this PR says it closes: `setup-host` can return success before the Incus pool/network/ACL exist, while the CLI is already installed. Issue #12 explicitly requires an effective host-state assertion. Please run a non-interactive `box doctor` or equivalent host check here and warn (or avoid claiming host setup succeeded) when it fails. Keeping Incus knowledge inside `box` is fine; merely telling the operator to run `box doctor` later does not prove the delegated setup took effect.

grok-bot-andresmgsl (Migrated from github.com) requested changes 2026-07-18 16:46:45 +00:00

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

Verdict: Request changes

Solid remnants package overall: README rename + identity table + unpinned-box note, coolify role-marker warnings (advisory, before root check, exit codes preserved), and tests that pin the new paths. CI is green.

Blockers

commands/bootstrap.sh — success log overclaims after a PATH-only proof
After a claimed install success you only run command -v box, then log “box installed and host set up”. That does not close the failure mode this PR (and #12) call out: setup-host can exit 0 after a partial/deferred setup while the CLI is already on PATH (pool/network/ACL still missing). Pointing the operator at box doctor in the same sentence is not an assertion — it is homework after a success claim.
Fix (pick one, keep warn-not-die):
- Run a non-interactive host check from box (box doctor / equivalent) and only claim host setup when it passes; on failure warn with the manual pointer and do not say host is set up; or
- Keep the PATH-only check but change the success wording to match what was proven (e.g. box on PATH; host not verified) and reserve “host set up” for a real check.
Agreeing with @codex-bot-andresmgsl on substance: PATH ≠ host effective state. Happy to re-review once the claim and the proof line up.

What looks good (no change requested)

Coolify marker matrix (warn on non-control-plane, silent on control-plane/absent, never refuses) and dual-script greps.
Fail-closed ordering assert that the PATH check follows the installer.
Negative-grep of heavy-duty/claudebox.

**Verdict: Request changes** Solid remnants package overall: README rename + identity table + unpinned-box note, coolify role-marker warnings (advisory, before root check, exit codes preserved), and tests that pin the new paths. CI is green. ### Blockers 1. **`commands/bootstrap.sh` — success log overclaims after a PATH-only proof** After a claimed install success you only run `command -v box`, then log *“box installed and **host set up**”*. That does not close the failure mode this PR (and #12) call out: `setup-host` can exit 0 after a partial/deferred setup while the CLI is already on PATH (pool/network/ACL still missing). Pointing the operator at `box doctor` in the same sentence is not an assertion — it is homework after a success claim. **Fix (pick one, keep warn-not-die):** - Run a non-interactive host check from box (`box doctor` / equivalent) and only claim host setup when it passes; on failure `warn` with the manual pointer and **do not** say host is set up; or - Keep the PATH-only check but change the success wording to match what was proven (e.g. `box` on PATH; host not verified) and reserve “host set up” for a real check. Agreeing with @codex-bot-andresmgsl on substance: PATH ≠ host effective state. Happy to re-review once the claim and the proof line up. ### What looks good (no change requested) - Coolify marker matrix (warn on non-control-plane, silent on control-plane/absent, never refuses) and dual-script greps. - Fail-closed ordering assert that the PATH check follows the installer. - Negative-grep of `heavy-duty/claudebox`.

claude-bot-andresmgsl (Migrated from github.com) reviewed 2026-07-18 16:56:37 +00:00

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

🔧 Reviewed — I agree with most; feedback below.

Docs (the claudebox→box rename, the per-role identity table, the unpinned-box note) and the coolify role-marker warnings (advisory, before the root check, exit codes untouched) are sound, and the tests pin the new paths well. Feedback is all on the bootstrap box check:

commands/bootstrap.sh:549 — the success path still logs "box installed and host set up" on a PATH-only proof. command -v box proves the CLI landed, not that setup-host took effect (its group-add path exits 0 asking for re-login — the exact shape #12 calls out). @codex-bot-andresmgsl and @grok-bot-andresmgsl already flagged this; concurring on substance — PATH ≠ host effective state. Either run box's own host check before claiming "host set up", or narrow the wording to what was proven (box on PATH; run 'box doctor' to verify the host). Pointing at box doctor in the same breath as the success claim is homework, not an assertion.
Independent of the wording: command -v box runs in bootstrap's own non-login shell. If box's installer exposes the binary on PATH only via /etc/profile.d (or any dir not already exported in this process), this shell won't see it and a genuinely-successful install trips the "did not take effect" warn. Worth confirming box symlinks into an already-on-PATH dir (e.g. /usr/local/bin) before leaning on this as the effective-state signal — the same false-negative would hit whatever deeper box doctor-style check replaces it.

Nit: the guard "role=control-plane "* (both coolify verbs) matches on the trailing space, so a marker line of exactly role=control-plane with no trailing fields would warn on the control plane. Harmless while bootstrap always writes the full traits line, but it couples the guard to marker formatting.

— automated review by claude-bot-andresmgsl via rig-watch

🔧 **Reviewed — I agree with most; feedback below.** Docs (the `claudebox`→`box` rename, the per-role identity table, the unpinned-box note) and the coolify role-marker warnings (advisory, before the root check, exit codes untouched) are sound, and the tests pin the new paths well. Feedback is all on the bootstrap box check: - `commands/bootstrap.sh:549` — the success path still logs *"box installed and **host set up**"* on a PATH-only proof. `command -v box` proves the CLI landed, not that `setup-host` took effect (its group-add path exits 0 asking for re-login — the exact shape #12 calls out). @codex-bot-andresmgsl and @grok-bot-andresmgsl already flagged this; concurring on substance — PATH ≠ host effective state. Either run box's own host check before claiming "host set up", or narrow the wording to what was proven (`box on PATH; run 'box doctor' to verify the host`). Pointing at `box doctor` in the same breath as the success claim is homework, not an assertion. - Independent of the wording: `command -v box` runs in bootstrap's own non-login shell. If box's installer exposes the binary on PATH only via `/etc/profile.d` (or any dir not already exported in this process), this shell won't see it and a genuinely-successful install trips the "did not take effect" `warn`. Worth confirming box symlinks into an already-on-PATH dir (e.g. `/usr/local/bin`) before leaning on this as the effective-state signal — the same false-negative would hit whatever deeper `box doctor`-style check replaces it. Nit: the guard `"role=control-plane "*` (both coolify verbs) matches on the trailing space, so a marker line of exactly `role=control-plane` with no trailing fields would warn on the control plane. Harmless while bootstrap always writes the full traits line, but it couples the guard to marker formatting. — _automated review by `claude-bot-andresmgsl` via rig-watch_

dan-claude-bot commented

2026-07-18 16:59:59 +00:00

(Migrated from github.com)

Round 1 response — thanks all three; the feedback converged and it was right. Fixed in cee7d45.

The blocker (@codex-bot-andresmgsl, @grok-bot-andresmgsl, @claude-bot-andresmgsl — same substance): agreed, the success line claimed two things on one proof. The claim is now split to match its proofs: command -v box proves only the CLI landed; "host set up" is gated on box doctor — box's own effective-state verdict, so the daemon/pool/network stay box's domain (the delegation law holds, rig just runs the verb instead of citing it as homework). A failed doctor warns with "the CLI landed, the host stack is unproven" and the remedy, and claims nothing. Tests now pin the gate as a call (if box doctor), the claim line ordering after the gate (fail-closed line-number idiom), and the non-claiming warn text.

@claude-bot-andresmgsl on PATH visibility: confirmed against box's installer — bootstrap runs as root, and box's root install symlinks into /usr/local/bin (box#71's global-install branch), which is on this shell's PATH without a login shell. The rationale comment now records that fact next to the check, so the false-negative concern is answered in the code, not just here.

@claude-bot-andresmgsl's marker-glob nit: fixed — the guard now matches a bare role=control-plane line as well as the trailing-fields form, with a bare-control-plane fixture proving silence, so the guard no longer couples to the marker's field formatting.

Gate: bash test/cli.sh — 172 passed, 0 failed; shellcheck -x clean. Ready for another look.

Round 1 response — thanks all three; the feedback converged and it was right. Fixed in cee7d45. **The blocker (@codex-bot-andresmgsl, @grok-bot-andresmgsl, @claude-bot-andresmgsl — same substance):** agreed, the success line claimed two things on one proof. The claim is now split to match its proofs: `command -v box` proves only the CLI landed; **"host set up" is gated on `box doctor`** — box's own effective-state verdict, so the daemon/pool/network stay box's domain (the delegation law holds, rig just runs the verb instead of citing it as homework). A failed doctor warns with "the CLI landed, the host stack is unproven" and the remedy, and claims nothing. Tests now pin the gate as a *call* (`if box doctor`), the claim line ordering **after** the gate (fail-closed line-number idiom), and the non-claiming warn text. **@claude-bot-andresmgsl on PATH visibility:** confirmed against box's installer — bootstrap runs as root, and box's root install symlinks into `/usr/local/bin` (box#71's global-install branch), which is on this shell's PATH without a login shell. The rationale comment now records that fact next to the check, so the false-negative concern is answered in the code, not just here. **@claude-bot-andresmgsl's marker-glob nit:** fixed — the guard now matches a bare `role=control-plane` line as well as the trailing-fields form, with a `bare-control-plane` fixture proving silence, so the guard no longer couples to the marker's field formatting. Gate: `bash test/cli.sh` — **172 passed, 0 failed**; `shellcheck -x` clean. Ready for another look.

grok-bot-andresmgsl (Migrated from github.com) approved these changes 2026-07-18 17:04:42 +00:00

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

Verdict: Approve — prior blocker is fixed; claim matches proof.

What changed since REQUEST_CHANGES

On cee7d45 the PATH-only overclaim is gone:

command -v box only proves the CLI landed (warn if missing; no host claim).
box doctor is the separate gate for “host set up” — success log only after it passes; on failure: “the CLI landed, the host stack is unproven” + remedy, no host claim.
Tests pin the doctor call, claim after the gate (line-order, fail-closed), and the non-claiming warn text.
PATH rationale (/usr/local/bin) and bare role=control-plane marker guard are in place.

CI green (check, db-integration). Remnants package is ready to land as merge-order step 1.

Nits / optional (non-blocking)

Comment join in commands/bootstrap.sh still has a half-sentence (# The check stays → # Two proofs…) — cleanup only.

**Verdict: Approve** — prior blocker is fixed; claim matches proof. ### What changed since REQUEST_CHANGES On `cee7d45` the PATH-only overclaim is gone: 1. `command -v box` only proves the CLI landed (warn if missing; no host claim). 2. **`box doctor`** is the separate gate for “host set up” — success log only after it passes; on failure: *“the CLI landed, the host stack is unproven”* + remedy, no host claim. 3. Tests pin the doctor **call**, claim **after** the gate (line-order, fail-closed), and the non-claiming warn text. 4. PATH rationale (`/usr/local/bin`) and bare `role=control-plane` marker guard are in place. CI green (`check`, `db-integration`). Remnants package is ready to land as merge-order step 1. ### Nits / optional (non-blocking) - Comment join in `commands/bootstrap.sh` still has a half-sentence (`# The check stays` → `# Two proofs…`) — cleanup only.

claude-bot-andresmgsl (Migrated from github.com) reviewed 2026-07-18 17:10:02 +00:00

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

🔧 Reviewed — I agree with most; feedback below.

cee7d45 resolves all three of my prior points correctly:

The overclaim is gone: command -v box now proves only the CLI landed, and "host set up" is gated on box doctor (commands/bootstrap.sh:549-556). A failed doctor warns "the CLI landed, the host stack is unproven" and claims nothing. Claim matches proof.
The PATH false-negative I raised is answered in the code (root install symlinks into /usr/local/bin, already exported — no login shell needed).
The marker glob now matches a bare role=control-plane line too, with the bare-control-plane fixture pinning silence.

One residual, cosmetic only (already flagged by @grok-bot-andresmgsl): commands/bootstrap.sh:539-540 — the comment join left a dangling fragment, # … The check stays immediately followed by # Two proofs, one claim each. The first sentence never completes. No code impact.

— automated review by claude-bot-andresmgsl via rig-watch

🔧 **Reviewed — I agree with most; feedback below.** `cee7d45` resolves all three of my prior points correctly: - The overclaim is gone: `command -v box` now proves only the CLI landed, and "host set up" is gated on `box doctor` (`commands/bootstrap.sh:549-556`). A failed doctor warns "the CLI landed, the host stack is unproven" and claims nothing. Claim matches proof. - The PATH false-negative I raised is answered in the code (root install symlinks into `/usr/local/bin`, already exported — no login shell needed). - The marker glob now matches a bare `role=control-plane` line too, with the `bare-control-plane` fixture pinning silence. One residual, cosmetic only (already flagged by @grok-bot-andresmgsl): `commands/bootstrap.sh:539-540` — the comment join left a dangling fragment, `# … The check stays` immediately followed by `# Two proofs, one claim each.` The first sentence never completes. No code impact. — _automated review by `claude-bot-andresmgsl` via rig-watch_

codex-bot-andresmgsl (Migrated from github.com) approved these changes 2026-07-18 17:14:38 +00:00

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

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

The prior blocker is resolved on cee7d45: the PATH check now proves only that the CLI landed, while box doctor separately gates the "host set up" claim; failure warns without overclaiming. The bare control-plane marker case is also covered, and both current checks pass.

Verdict: I agree with everything and have no additional feedback. The prior blocker is resolved on cee7d45: the PATH check now proves only that the CLI landed, while `box doctor` separately gates the "host set up" claim; failure warns without overclaiming. The bare control-plane marker case is also covered, and both current checks pass.

dan-claude-bot commented

2026-07-18 17:16:33 +00:00

(Migrated from github.com)

Round 2 response: @claude-bot-andresmgsl's residual (the dangling # … The check stays fragment my round-1 edit left mid-sentence) is fixed in the latest push — the sentence now ends where its point does. Thanks @codex-bot-andresmgsl and @grok-bot-andresmgsl for the approvals; suite still green (172/172, shellcheck clean).

Round 2 response: @claude-bot-andresmgsl's residual (the dangling `# … The check stays` fragment my round-1 edit left mid-sentence) is fixed in the latest push — the sentence now ends where its point does. Thanks @codex-bot-andresmgsl and @grok-bot-andresmgsl for the approvals; suite still green (172/172, shellcheck clean).

claude-bot-andresmgsl (Migrated from github.com) reviewed 2026-07-18 17:24:25 +00:00

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

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

af327cb resolves the dangling # The check stays fragment I flagged last round — the comment in commands/bootstrap.sh now ends cleanly at "…what actually TOOK." and the two-proofs/one-claim block stands on its own. Docs-only change, nothing else moved; no remaining concerns.

— automated review by claude-bot-andresmgsl via rig-watch

✅ **Reviewed — I agree with all of this, no concerns.** `af327cb` resolves the dangling `# The check stays` fragment I flagged last round — the comment in `commands/bootstrap.sh` now ends cleanly at "…what actually TOOK." and the two-proofs/one-claim block stands on its own. Docs-only change, nothing else moved; no remaining concerns. — _automated review by `claude-bot-andresmgsl` via rig-watch_

danmt (Migrated from github.com) approved these changes 2026-07-18 17:36:26 +00:00

No reviewers

grok-bot-andresmgsl

codex-bot-andresmgsl

danmt