chore: remove agent-rig workflow artifacts (release-gates/, private plan docs) (#4632)

The release-gates/ directory (16 files) and two docs/plans/ files
(schema-skew-guard.md, reference-aware-prune.md) are internal artifacts
from a contributor's personal multi-agent development rig: release-gate
checklists and PM handoff plans referencing bead IDs (be-*) from a
private tracker that does not exist in this repo. They arrived as
passengers on otherwise-valuable contribution PRs between April and
July 2026 and have no readers here - the user-facing documentation for
those features landed in --help text and README via the same PRs.

Removing them keeps the tree free of files that reference untraceable
issue IDs and confuse both human readers and coding agents that treat
docs/plans/ as authoritative. Content remains available in git history.

Agent-Signature: claude-fable-5-high on behalf of matt wilkie
This commit is contained in:
matt wilkie 2026-07-07 09:54:01 -07:00 • committed by GitHub
commit 187ba85f3c
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
19 changed files with 0 additions and 1135 deletions

View file

@ -1,59 +0,0 @@
# Plan: Reference-Aware `bd prune` (be-zej8mz)
**PM:** beads/pm
**Date:** 2026-05-17
**Architect spec:** be-5sn
**Designer spec:** be-zej8mz (description field)
**Status:** Decomposed → builder handoff
---
## Goal
`bd prune` gains a reference-protection layer: closed beads whose ID appears in any open bead's description, notes, or comments are silently skipped. A new `--ignore-references` flag provides the escape hatch.
This closes the incident from 2026-05-01 where architect ADR beads (be-08pl, be-eei) were pruned despite being cited by open verification beads.
---
## Child beads
| ID | Title | Label | Target |
|----|-------|-------|--------|
| be-jewoem | Implement reference-aware bd prune core logic | ready-to-build | builder |
| be-u2mw2x | Update bd prune --help and README for reference-aware skip | ready-to-build | builder |
| be-0wt833 | Bench: bd prune reference scan under 5s on 10K-open-bead fixture | ready-to-build | builder (blocked on be-jewoem) |
---
## Dependency graph
```
be-zej8mz (parent)
├── be-jewoem core impl ← unblocked, ready to build
├── be-u2mw2x docs update ← unblocked, ready to build
└── be-0wt833 bench ← blocked until be-jewoem merges
```
---
## Key implementation notes for builder
- **Use `Statuses` filter** (not `ExcludeStatus`) when fetching non-closed beads — PG has a coverage gap on `ExcludeStatus` (be-jdeief). Statuses: open, in_progress, blocked, deferred, pinned.
- **`--ignore-references` on `pruneCmd` only** — `bd purge --ignore-references` must return "unknown flag".
- **`referenced_skipped` always in JSON** (even 0) so orchestrators can branch; `referenced_ids_sample` omitted when 0.
- **Human output**: ≤5 IDs in sample; JSON: ≤100 IDs in sample.
- **No new colors** — reuse `ui.MutedStyle`, `ui.IDStyle`, `ui.CommandStyle` from existing palette.
- **Exact output copy** in be-zej8mz §4 (designer spec) — match it precisely.
---
## Acceptance summary
1. `bd prune` skips referenced closed beads by default.
2. `bd prune --ignore-references` deletes them.
3. `bd purge --ignore-references` → "unknown flag".
4. Integration test: 3-bead fixture (1 referenced, 1 pinned, 1 plain) → only plain deleted.
5. Bench: 10K-bead fixture scan <5s (skippable with `testing.Short()`).
6. All output matches be-zej8mz §4 exact copy.
7. `go test ./...` passes; `golangci-lint` clean.

View file

@ -1,99 +0,0 @@
# Plan: Schema-Skew Guard Implementation
**Filed:** 2026-05-22
**PM:** beads/pm
**Root cause:** be-85qfg — stale bd binary accumulated ~25,000 errno-1105 errors over 16h
**Architecture:** be-blpg3
**Status:** In-flight
---
## Goal
Hard-fail at db-open time when the database schema is ahead of the binary's
embedded migrations, preventing cryptic SQL errors caused by stale binaries.
## Work Breakdown
### be-wwbsv — Implement schema-skew guard (P1, builder)
**What:** Core implementation of the schema-skew guard.
**Files:**
- `internal/storage/schema/schema.go` — SchemaSkewError type, IsSchemaSkewError helper, checkSchemaSkew (unexported), CheckForwardDrift (exported), call in MigrateUp
- `internal/storage/dolt/store.go` — CheckForwardDrift in ReadOnly branch of newServerMode
- `cmd/bd/main.go` — --ignore-schema-skew PersistentFlag, env-var set in PersistentPreRun, dedicated SchemaSkewError handler before FatalError
**Key UX decisions (from designer, be-wwbsv):**
- Dedicated `errors.As(err, &skewErr)` check before FatalError (avoids triple-prefix)
- SchemaSkewError.UserMessage() for human output, Error() for terse one-liner
- Singular/plural: "1 migration ahead" / "N migrations ahead"
- Warning line: `Warning: schema skew ignored — database (v{DBVersion}) is ahead of binary (v{BinaryVersion}); some queries may fail`
- JSON: extend buildJSONError with "schema_skew" subobject (current_version, required_version, delta)
**Acceptance criteria:**
- `bd list` / `bd ready` exits 1 with actionable SchemaSkewError when DB > binary
- BD_IGNORE_SCHEMA_SKEW=1 bypasses guard, prints Warning line to stderr
- --ignore-schema-skew flag sets env var in PersistentPreRun
- --json flag produces schema_skew subobject on stderr
- Fresh DB (currentVersion=0) and normal upgrade path unaffected
- bd migrate and bd init excluded from guard
### be-x0rl6 — Document schema-skew guard (P2, builder)
**What:** Documentation additions (no new files).
**Files:**
- `cmd/bd/main.go` — flag description for --ignore-schema-skew (exact copy in be-x0rl6 §2a)
- `README.md` — ### Schema Version Guard section after ### Backup & Migration (exact copy in be-x0rl6 §3b)
- `CHANGELOG.md` — entry under [Unreleased] → Added (exact copy in be-x0rl6 §4b)
**Acceptance criteria:**
- `bd --help` shows --ignore-schema-skew with exact description text
- README renders ### Schema Version Guard section with correct code block (using 4-backtick or ~~~ fence)
- CHANGELOG entry present under [Unreleased] → Added
### be-cdhvj — Tests: schema-skew guard (P1, validator)
**Blocked on:** be-wwbsv
**What:** Unit, integration, and escape-hatch test coverage.
**Acceptance criteria:**
- Unit: checkSchemaSkew mock for version=0/equal/+1/+3, BD_IGNORE_SCHEMA_SKEW path
- Error copy exact-match assertions: Error(), UserMessage(), warning line, EscapeHint()
- JSON --json flag: schema_skew subobject present with correct field types
- Integration: dolt.New at N+1 returns *SchemaSkewError; + BD_IGNORE_SCHEMA_SKEW succeeds
- ReadOnly CheckForwardDrift: same integration scenarios
- bd init guard no-op (fresh DB)
- IsSchemaSkewError distinguishes skew from other errors
---
## Dependency Graph
```
be-wwbsv (impl) ──blocks──> be-cdhvj (tests)
be-x0rl6 (docs) ── independent
```
Builder ships be-wwbsv and be-x0rl6 in one PR. Validator writes be-cdhvj after.
---
## Routing
| Bead | Agent | Label |
|------|-------|-------|
| be-wwbsv | beads/builder | ready-to-build |
| be-x0rl6 | beads/builder | ready-to-build |
| be-cdhvj | beads/validator | needs-tests |
---
## Guardrails (carry forward from be-blpg3)
- BD_IGNORE_SCHEMA_SKEW is the ONLY bypass — no ReadOnly=true bypass
- checkSchemaSkew returns nil on currentVersion=0 — do not break bd init
- CheckForwardDrift called only in ReadOnly branch (write path gets it via MigrateUp)
- Do NOT add guard to bd migrate command
- Error message in README must match SchemaSkewError.UserMessage() exactly — update both in same PR

View file

@ -1,39 +0,0 @@
# Release Gate: bd-3852 — orphaned children doctor check
**PR**: https://github.com/gastownhall/beads/pull/4053
**Branch**: `fix/bd-3852-orphaned-children-check` (gastownhall/beads)
**Date**: 2026-05-21
**Deployer**: beads/deployer
## Gate Result: PASS
| # | Criterion | Evidence | Result |
|---|-----------|----------|--------|
| 1 | Review PASS present | be-qodg: PASS — 2 LOW notes, no blockers | ✅ PASS |
| 2 | Acceptance criteria met | See below | ✅ PASS |
| 3 | Tests pass | CI run 26173640486 — all 41 checks PASS (ubuntu, macOS, Windows smoke, Embedded Dolt cmd 1–20, Storage, lint, fmt, doc freshness, upgrade smokes) | ✅ PASS |
| 4 | No high-severity findings | be-qodg: "None blocking" — LOW-1 cosmetic `%d+` unconditional; LOW-2 no functional impact | ✅ PASS |
| 5 | Final branch is clean | `git status` — clean; only rig artifacts untracked | ✅ PASS |
| 6 | Branch diverges cleanly from main | 1 commit ahead of `origin/main`; merge state CLEAN | ✅ PASS |
## Acceptance Criteria Verification (bd-3852)
| Criterion | Evidence | Result |
|-----------|----------|--------|
| `checkOrphanedChildren(db)` added to `RunDeepValidation` | `result.OrphanedChildren = checkOrphanedChildren(db)` at line 121; wired into `AllChecks` at line 122 | ✅ |
| `OrphanedChildren DoctorCheck` field in `DeepValidationResult` | `OrphanedChildren DoctorCheck \`json:"orphaned_children"\`` at line 22 | ✅ |
| Query uses `SUBSTRING_INDEX` to find children whose parent prefix is not in issues | SQL uses `SUBSTRING_INDEX(id, '.', 1) NOT IN (SELECT id FROM issues) LIMIT 10` | ✅ |
| Reports as `StatusWarning`, not `StatusError` | `check.Status = StatusWarning` on orphans found path | ✅ |
| Shows up to 3 example IDs | `orphans[:min(3, len(orphans))]` in Detail field | ✅ |
| Fix hint provided | `check.Fix = "Delete orphaned issues with 'bd admin delete <id>' or re-create the parent"` | ✅ |
| `rows.Err()` checked | Present after row scan loop | ✅ |
| `defer rows.Close()` in place | Present | ✅ |
| Dolt-backend guard prevents `SUBSTRING_INDEX` on SQLite | `RunDeepValidation` returns early if `backend != BackendDolt` (reviewer confirmed) | ✅ |
**Note on LOW-1**: Message says `"Found %d+ child issues whose parent no longer exists"` — the `+` is unconditional even when count < LIMIT (e.g., "Found 3+ child issues" is misleading when exactly 3 exist). This is cosmetic and non-blocking per reviewer verdict.
## Commits
| SHA | Description |
|-----|-------------|
| `94298a169` | feat(doctor): detect orphaned child issues in deep validation (bd-3852) |

View file

@ -1,68 +0,0 @@
# Release gate — be-07y (be-035 close authority guard)
- **Bead:** be-07y (review bead for fix of be-035, P1 silent data loss)
- **Commit shipped:** 3dd4375c (cherry-pick of 7a925e304 from `users/jaword/postgres-backend`)
- **Branch:** `release/be-07y-gate` off `origin/main`
- **Evaluated:** 2026-05-05 by beads/deployer
## Scope note
The builder committed `7a925e304` on top of `users/jaword/postgres-backend`, a
branch that carries ~10 unrelated postgres-backend commits explicitly parked as
LOCAL-ONLY (be-7yt, be-b0h, be-ry0, be-2oq, be-xz4, be-y8g, be-6fk.x, …).
Shipping the builder's branch directly would drag those parked commits into
the PR. To keep the unit-of-review tight, this gate cuts a clean branch off
`origin/main` and cherry-picks only `7a925e304`. Resulting commit `3dd4375c`
introduces a delta byte-identical to the reviewed commit (verified via
`diff <(git show 7a925e304 -- <files>) <(git show HEAD -- <files>)` — only
commit/blob SHAs differ; patch hunks are identical).
## Gate criteria
| # | Criterion | Verdict | Evidence |
|---|-----------|---------|----------|
| 1 | Review PASS present | **PASS** | beads/reviewer recorded `PASS` verdict in be-07y notes (single-pass; gemini second-pass disabled). Three non-blocking findings (LOW-1 filed as be-po9, LOW-2 observability, INFO-1 scope choice). |
| 2 | Acceptance criteria met | **PASS** | All four be-035 criteria covered by the test set. (1) Authority-mismatch reproducer in `TestEmbeddedClose/close_assignee_mismatch_refuses_without_force` asserts non-zero exit, stderr names actor/assignee, bead remains open. (2) Concurrent two-actor scenario covered by claim/close pair in same sub-test. (3) Refused close exits non-zero (`os.Exit(1)` at `cmd/bd/close.go:277` when `closedCount==0`). (4) Clear stderr message format `cannot close %s: assignee is %q, actor is %q; reclaim or use --force to override`. |
| 3 | Tests pass | **PASS** | See "Tests run on release branch" below. Three flaky failures observed under full-suite parallelism are pre-existing baseline (reproduce on plain `origin/main`); not regressions. |
| 4 | No high-severity review findings open | **PASS** | 0 HIGH findings. LOW-1 → follow-up be-po9 (P2). LOW-2 + INFO-1 advisory. |
| 5 | Final branch is clean | **PASS** | `git status` on `release/be-07y-gate` shows nothing except worktree-scaffolding untracked paths (`.gc/`, `.gitkeep`). |
| 6 | Branch diverges cleanly from main | **PASS** | Branch cut fresh from `origin/main` via `git checkout -B release/be-07y-gate origin/main`; `git cherry-pick 7a925e304` applied with auto-merge on 4 files (context-only; verified delta is byte-identical to the original commit). |
## Tests run on release branch
| Test | Result | Notes |
|------|--------|-------|
| `TestAssigneeMatches` (`internal/validation/`, plain go test) | PASS 0.003s | All 6 sub-tests: nil_issue, unassigned, matching, mismatched_no_force, mismatched_with_force, empty_actor. |
| `TestValidateIssueClosable` (`cmd/bd/`, plain go test) | PASS 0.159s | Validator chain wiring with new `actor` parameter. |
| `TestEmbeddedClose` (`cmd/bd/`, `BEADS_TEST_EMBEDDED_DOLT=1`, `-tags 'cgo dolt_only'`) | PASS 65.05s | All 35 sub-tests, including the 4 new ones: `close_assignee_mismatch_refuses_without_force`, `close_assignee_mismatch_with_force`, `close_same_actor_succeeds`, `close_unassigned_bead_succeeds`. |
| `TestEmbeddedCloseConcurrent` (same tags) | PASS 65.57s | 50 issues × 10 concurrent workers; 50 in DB. |
| `go vet -tags gms_pure_go ./internal/validation/... ./cmd/bd/` | clean | No output. |
| `golangci-lint run --new-from-rev=origin/main ./...` | clean | 0 new issues. |
| `make test` (full suite) | 3 flaky failures, all baseline | `TestPullWithAutoResolve_BranchTrackingFallback` (nil DB panic in cleanup, full-suite contention only — passes in isolation), `TestResolvePartialID_Wisp/wisp_prefix_with_hash` (passes in isolation), `TestApplyConfigDefaults_ProductionFallback` (port-resolution flake — reproduces on plain `origin/main` with no cherry-pick applied). None touch close.go, validation/issue.go, or any path in the be-07y delta. |
## Findings tracked from review
**LOW-1 (defense-in-depth — follow-up filed):** `bd update --status=closed`
bypasses the new `AssigneeMatches` guard because it routes through
`issueStore.UpdateIssue` directly without `validateIssueClosable`. Same
silent-data-loss class as be-035 via a different command surface. Builder
scoped fix to `bd close` per the be-035 ticket; reviewer filed **be-po9**
(P2 bug, `discovered-from be-07y`) for the parallel surface. Not a release
blocker.
**LOW-2 (observability — advisory):** Refused authority attempts log to
stderr but produce no `audit.LogFieldChange` entry. Optional improvement
for automation-heavy environments. Not in be-035 acceptance, not a
blocker.
**INFO-1 (scope choice — accepted):** be-035 mentioned "owner/admin
override"; the fix substitutes `--force` for that override. The narrowing
is sensible given bd's actual auth model (no role-based authorization;
`Owner` is a free-form CV-chain field). `--force` is the established
override pattern (NotPinned, gate-satisfaction, blocker bypass). Builder
flagged the substitution explicitly; reviewer accepted.
## Verdict
**PASS** — push to `fork` (origin is locked for quad341; `fork =
quad341/beads`), open PR against `gastownhall/beads:main`.

View file

@ -1,27 +0,0 @@
# Release Gate: be-0x7ztf — fix flaky TestProxy_IdleTimeout_Fires
**Branch:** `deploy/be-0x7ztf-proxy-listen-wait` (cherry-pick of d647f8837 onto origin/main)
**Date:** 2026-05-16
**Deployer:** beads/deployer
## Gate Checklist
| # | Criterion | Result | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | **PASS** | be-lw9lkj: "Review verdict: PASS" — no findings, gofmt clean, all acceptance criteria met |
| 2 | Acceptance criteria met | **PASS** | server_test.go:26 reads `listenWait = 10 * time.Second`; `go test ./internal/storage/dbproxy/proxy/...` exits 0; `gofmt -l` clean; no other lines changed |
| 3 | Tests pass | **PASS** | `go test ./internal/storage/dbproxy/proxy/... -count=1`: `ok` 6.012s |
| 4 | No high-severity findings open | **PASS** | Review table: 0 findings (— / — / — / —) |
| 5 | Final branch is clean | **PASS** | One cherry-pick commit; `git status` clean |
| 6 | Branch diverges cleanly from main | **PASS** | Clean cherry-pick of d647f8837 onto origin/main — zero conflicts |
**Overall: PASS**
## Branch Note
The builder's feature branch (`fix/be-0x7ztf-proxy-listen-wait`) was based on
`feat/be-fqjs3v-bdd-daemon-foundation` and carried 5 unrelated upstream commits.
The deploy branch is a clean cherry-pick of just the fix commit (d647f8837) onto
`origin/main`, since `internal/storage/dbproxy/proxy/server_test.go` already
exists on main and the patch applies without conflict. The reviewer reviewed the
one-line change; the deploy branch carries exactly that change.

View file

@ -1,68 +0,0 @@
# Release gate — be-3lch (initConfigForTest auto-opt-out of repo config)
- **Bead:** be-veon (review of build be-3lch / architect be-lspm)
- **Commit shipped:** `30c1ad1c7` (cherry-pick of `6858f98fc` from `fix/be-0d4-postgres-init-guards`)
- **Branch:** `release/be-3lch` off `origin/main`
- **Evaluated:** 2026-05-12 by beads/deployer
## Scope note
The source commit `6858f98fc` is a **+5 LOC test-only** change to
`cmd/bd/test_helpers_pure_test.go`. The helper now calls
`t.Setenv("BEADS_TEST_IGNORE_REPO_CONFIG", "1")` as its first statement,
so every test that uses `initConfigForTest(t)` gets a clean viper by
default (no silent inheritance of the dogfooded `<repo>/.beads/config.yaml`).
A 3-line doc comment update explains the opt-out.
**Relationship to be-sm6v:** be-sm6v (release/be-sm6v, open PR) adds the
explicit per-test `t.Setenv` + `t.Chdir(t.TempDir())` defense in
`where_cgo_test.go` to preserve bisect signal for the original be-hple
failure. be-3lch generalizes that opt-out into the shared helper so all
seven `initConfigForTest` callers get it for free. The two PRs are
designed to land in order (be-sm6v first); regardless of merge order,
neither change touches production code.
**Architect's deviation handled:** design called for a lockstep edit in
two helper files. be-5t00 (commit `b589f9b63`) already removed the
cgo-tagged duplicate `test_helpers_test.go::initConfigForTest`; the
helper now lives only in `test_helpers_pure_test.go` and a single-file
edit satisfies the design intent.
## Gate criteria
| # | Criterion | Verdict | Evidence |
|---|-----------|---------|----------|
| 1 | Review PASS present | **PASS** | beads/reviewer recorded `Verdict: PASS` in be-veon notes (gm-81uq8l, 2026-05-12 08:21). Reviewer verified all 4 architect acceptance criteria, all 7 caller paths, and confirmed no production code touched. Pre-existing flake (`TestResolveWhereBeadsDir_ReturnsEmptyWithoutWorkspace`) reproduces on parent commit, doesn't touch `initConfigForTest`. |
| 2 | Acceptance criteria met | **PASS** | All 4 architect criteria satisfied: (1) `./scripts/test.sh ./cmd/bd/` zero new regressions vs origin/main baseline (16 pre-existing failures, identical on both); (2) `t.Setenv("BEADS_TEST_IGNORE_REPO_CONFIG", "1")` placed as first statement after `t.Helper()`, before `config.ResetForTesting()` at `cmd/bd/test_helpers_pure_test.go:91`; (3) doc comment above helper updated with the 3-line opt-out hint (lines 85-87); (4) audit gate `TestWhereCommand_UsesConfigPrefixFromSelectedDB` PASSES in isolation. |
| 3 | Tests pass | **PASS** | Full `./scripts/test.sh ./cmd/bd/` on `release/be-3lch`: 16 failures, **identical set** to origin/main baseline (verified by running same command on `da73b7511`). Zero new regressions introduced. See "Tests run" below. |
| 4 | No high-severity review findings open | **PASS** | 0 HIGH findings. Reviewer notes "Findings: none blocking. Releasing to deployer." |
| 5 | Final branch is clean | **PASS** | `git status` on `release/be-3lch` shows nothing modified; only worktree-scaffolding untracked paths (`.gc/`, `.gitkeep`) that are never staged. |
| 6 | Branch diverges cleanly from main | **PASS** | Branch cut via `git checkout -B release/be-3lch da73b7511` (origin/main HEAD at evaluation time); `git cherry-pick 6858f98fc` applied with zero conflicts; `git log origin/main..HEAD` = 1 commit (plus this gate commit). |
## Tests run on release branch
| Test | Result | Notes |
|------|--------|-------|
| `TMPDIR=~/.gotmp GOTMPDIR=~/.gotmp ./scripts/test.sh ./cmd/bd/` (release/be-3lch) | 16 fails | Same failures as origin/main baseline. Failing tests do NOT call `initConfigForTest` — they are pre-existing state-leakage tests touching `$HOME/.beads` and worktree-hooks plumbing. |
| `TMPDIR=~/.gotmp GOTMPDIR=~/.gotmp ./scripts/test.sh ./cmd/bd/` (origin/main baseline `da73b7511`) | 16 fails | Identical failure set, verifying no regression from the be-3lch change. |
| `go test -tags gms_pure_go -run TestWhereCommand_UsesConfigPrefixFromSelectedDB -v ./cmd/bd/` (audit gate, architect criterion #4) | PASS 0.00s | Audit gate confirmed against the new helper. |
## Environment caveats (informational, not gate-blocking)
When run **in isolation** from a gc-rig worktree (where the moduleRoot is
NOT the beads source tree), `TestWhereCommand_ReadsPrefixFromEmbeddedStore`
still reads `issue-prefix: be` via viper's worktree-fallback probe —
which `BEADS_TEST_IGNORE_REPO_CONFIG=1` does not suppress because it only
ignores `<moduleRoot>/.beads/config.yaml`, not the cousin beads dir found
via `git rev-parse --git-common-dir`. This is exactly the trap be-sm6v's
`t.Chdir(t.TempDir())` was designed to close (see be-sm6v gate). In a
clean CI checkout (moduleRoot == beads tree), the helper-level fix from
be-3lch is sufficient. The failing-in-isolation test is also not visible
in the full suite because an earlier test (`TestInit_WithBEADS_DIR_DoltBackend`)
terminates the test binary before the cgo tests run — this is a
pre-existing condition, identical on the baseline.
## Verdict
**PASS** — push to `fork` (origin is locked for quad341; `fork =
quad341/beads`), open PR against `gastownhall/beads:main`.

View file

@ -1,49 +0,0 @@
# Release Gate: be-7daa14 — feat(init): auto-configure contributor routing on fork detect
**PR**: https://github.com/gastownhall/beads/pull/4028
**Branch**: `fix/be-7daa14-fork-detection-output` (quad341/beads)
**Date**: 2026-05-19
**Deployer**: beads/deployer
**Maintainer update**: 2026-05-23, original PR branch updated with a maintainer
merge from current `upstream/main` plus follow-up fix/test commits. The merge
keeps the current disabled-by-default auto-export behavior from `main`, and the
follow-up fix places `autoConfigureForkContributor` before the initial
`store.Commit` / `store.Close` so routing config is written to an open store and
included in the initial Dolt commit.
## Gate Result: PASS
| # | Criterion | Evidence | Result |
|---|-----------|----------|--------|
| 1 | Review PASS present | be-rev-169eb4 PASS (2026-05-18) | ✅ PASS |
| 2 | Acceptance criteria met | See below | ✅ PASS |
| 3 | Tests pass | CI run 26009117468 all checks PASS; maintainer update adds `BEADS_TEST_EMBEDDED_DOLT=1 go test -tags gms_pure_go ./cmd/bd -run 'TestEmbeddedInit/fork_auto_contributor' -count=1` PASS | ✅ PASS |
| 4 | No high-severity findings | be-rev-169eb4: "None blocking" — N1/N2/N3 are non-blocking observations about output text deviations from the design mockup; design accepted them as semantically correct | ✅ PASS |
| 5 | Final branch is clean | `git status` — clean; only `.gc/` and `.gitkeep` untracked (rig artifacts) | ✅ PASS |
| 6 | Branch diverges cleanly from main | Original PR conflicted with current `main` around the auto-export block; maintainer merge resolved the conflict and preserved current `main` semantics | ✅ PASS |
## Acceptance Criteria Verification (be-7daa14 + be-0ccf34 design spec)
| Scenario | Expected output | Code path | Result |
|----------|-----------------|-----------|--------|
| Happy path (fork, first run) | `▶ Fork detected — configuring contributor routing` + `upstream: <url>` + 3–4 `✓` lines + opt-out hint | `autoConfigureForkContributor`: `isFork=true`, `existing=""`, `!quiet` block at function end | ✅ |
| Opt-out (`--role=maintainer` on fork) | `⚠ Fork detected (upstream: <url>) / Contributor routing skipped / To set up later: bd init --contributor` | `roleFlag == "maintainer"` branch | ✅ |
| Re-init (already configured) | `⚠ Fork detected (upstream: <url>) / already configured → <path> / Skipping auto-setup` | `existing != ""` branch | ✅ |
| CI / `--quiet` | Silent (no fork block) | All output guarded by `!quiet` | ✅ |
| No new lipgloss styles | Only `ui.RenderAccent`, `ui.RenderPass`, `ui.RenderWarn` used | Code review confirmed | ✅ |
| All existing init tests pass | CI run 26009117468 all PASS | CI evidence | ✅ |
| `bd init` wires `autoConfigureForkContributor` | Called in `init.go` before initial `store.Commit` / `store.Close` so store-backed config writes succeed | Maintainer update verified | ✅ |
| Suppress output in non-interactive mode | Second commit `32d3eb41e` adds `nonInteractive` guard | Code review confirmed | ✅ |
| Fork auto-routing regression coverage | Temp git fork with `upstream`, non-interactive quiet init, read back `routing.mode`, `routing.contributor`, `sync.remote`, and `beads.role` | `TestEmbeddedInit/fork_auto_contributor` | ✅ |
## Commits
| SHA | Description |
|-----|-------------|
| `0a5a16eaa` | feat(init): auto-configure contributor routing on fork detect (be-7daa14) |
| `32d3eb41e` | fix(init): suppress autoConfigureForkContributor output in non-interactive mode |
| `7038872a9` | chore: release gate PASS for be-7daa14 |
| `a60e62c7a` | Merge upstream/main into PR 4028 |
| `0772de4cb` | test(init): cover fork auto contributor routing |
| `979e039b4` | fix(init): run fork auto routing before store close |

View file

@ -1,71 +0,0 @@
# Release Gate: be-drsfc2 — schema: print migration name to stderr on apply
**Result: GATE PASS**
**Date:** 2026-06-13
**Deployer:** beads/deployer
**Branch under review:** `fix/be-drsfc2-migration-progress` (3 feature commits ahead of origin/main)
**Branch tip:** `875d586a4febf7946d26f2a864889a91f05560b3`
**Target:** origin/main (`e8ae7a291`)
**PR:** https://github.com/gastownhall/beads/pull/3914
**Deploy bead:** be-soq0
---
## Criteria Checklist
| # | Criterion | Result | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | **PASS** | beads/reviewer (Claude Sonnet 4.6) PASS at `875d586a4` — be-hlo7 bead notes, 2026-06-13. |
| 2 | Acceptance criteria met | **PASS** | See below. All criteria verified by reviewer against current tip. |
| 3 | Tests pass | **PASS** | `go test ./internal/storage/schema/...` → 62+ PASS, 0 FAIL (1.291s). CI Gate Required: PASS (both runs). See below. |
| 4 | No high-severity review findings open | **PASS** | Reviewer: "No blockers." One informational note (PR sequencing with #3919) is non-blocking. |
| 5 | Final branch is clean | **PASS** | `git status` clean. |
| 6 | Branch diverges cleanly from main | **PASS** | `git merge --no-commit --no-ff origin/main` → "Already up to date." Zero conflicts. MERGEABLE per GitHub. |
| 7 | Single feature theme | **PASS** | All commits address one behavior: print migration name to stderr when a TTY is attached. Simpler, terminal-gated sibling of PR #3919. |
---
## Criterion 2: Acceptance Criteria
| Spec | Code | Verdict |
|------|------|---------|
| Print migration name to stderr on apply | `fmt.Fprintf(stderr, "migrating schema: %s\n", ...)` per migration | ✅ |
| Only when stderr is a TTY (CI/pipe-friendly) | `defaultStderr()` uses `golang.org/x/term` to check; returns `io.Discard` when not a TTY | ✅ |
| `migrationSource` passed through `runMigrations` | Commit `81f1c7f79` adds `src migrationSource` param; previously always used `mainSource` | ✅ |
| Coverage: emission and line count | `TestRunMigrationsStderrOutput` verifies 1 line per migration | ✅ |
| Coverage: respects `src` argument | `TestRunMigrationsUsesProvidedSource` proves `runMigrations` uses the provided source | ✅ |
---
## Criterion 3: Test Run
```
go test -v -count=1 ./internal/storage/schema/...
...
=== RUN TestRunMigrationsStderrOutput
--- PASS: TestRunMigrationsStderrOutput (0.00s)
=== RUN TestRunMigrationsUsesProvidedSource
--- PASS: TestRunMigrationsUsesProvidedSource (0.00s)
PASS
ok github.com/steveyegge/beads/internal/storage/schema 1.291s
```
`go build ./internal/storage/schema/...` — PASS (via CI Build Artifacts: PASS)
`go vet ./internal/storage/schema/...` — PASS
CI Gate / Required — PASS (PR #3914, both runs: 27475151206 + 27475151232)
---
## Commits in PR
| Commit | Message |
|--------|---------|
| `8b939d27c` | schema: print migration name to stderr when running (human UX only) |
| `81f1c7f79` | fix(schema): pass migrationSource to runMigrations, not just mainSource |
| `875d586a4` | schema: suppress migration progress when stderr is not a TTY |
---
## Sequencing Note
PR #3919 (`rebase/be-ldr-2026-05-12`) also modifies the same `runMigrations` extraction site with a richer implementation (adds progress timing and large-rig warning). Both PRs independently PASS. Mayor/mpr must merge one first; the second will need a rebase. Recommended merge order: either works — #3914 is a simpler subset; #3919 is a superset. If #3919 merges first, #3914 may become redundant (superset already includes the name-to-stderr behavior).

View file

@ -1,62 +0,0 @@
# Release gate — be-eqw (WispIDSetInTx amortized wisp routing)
- **Bead:** be-eqw (review bead for build be-nu4.2.1 / ADR be-nu4 §4.D2)
- **Commit shipped:** 12ab3647 (cherry-pick of 61cfc45c from gc-builder-e35c0415a93c)
- **Branch:** `release/be-eqw` off `origin/main`
- **Evaluated:** 2026-04-23 by beads/deployer
## Scope note
The builder worktree `gc-builder-e35c0415a93c` carries two commits on top of
`origin/main`:
- `61cfc45c` — D2 build (be-nu4.2.1, reviewed PASS as be-eqw) — **shipped**
- `5023e0e1` — D3 build (be-nu4.3.2, review be-3ht still open) — **excluded**
This PR cuts a clean branch off `origin/main` with only the D2 commit, so the
un-reviewed D3 change does not ride along.
## Gate criteria
| # | Criterion | Verdict | Evidence |
|---|-----------|---------|----------|
| 1 | Review PASS present | **PASS** | reviewer-1 recorded `Verdict: PASS` in be-eqw notes (single-pass; gemini second-pass currently disabled). Findings F1 (medium scope-completeness), F2 (low benchmark-numbers) are advisory/follow-up, not blockers. |
| 2 | Acceptance criteria met | **PASS** | Reviewer trace table: hot-path callers make one wisp-id query per invocation ✓; mixed-ID routing hard gate ✓; `IsActiveWispInTx` retained for single-ID paths ✓; benchmark code at 1K/10K/50K with ≥25% wisp share ✓; no correlated EXISTS / recursive CTEs / UNION ALL ✓. Full-audit gap tracked as explicit follow-up (see §F1 below). |
| 3 | Tests pass | **PASS** | See "Tests run on release branch" below. |
| 4 | No high-severity review findings open | **PASS** | 0 HIGH findings. F1=medium, F2=low. |
| 5 | Final branch is clean | **PASS** | `git status` on `release/be-eqw` shows nothing except worktree-scaffolding untracked paths (`.gc/`, `.gitkeep`) that are never staged. |
| 6 | Branch diverges cleanly from main | **PASS** | Branch cut fresh from `origin/main` via `git checkout -B release/be-eqw origin/main`; `git cherry-pick 61cfc45c` applied with zero conflicts. |
## Tests run on release branch
| Test | Result | Notes |
|------|--------|-------|
| `go build ./...` | PASS | Go 1.26.2 via `GOTOOLCHAIN=auto`. |
| `go vet ./...` | clean | No output. |
| `gofmt -l` on 6 changed files | clean | No output. |
| `TestPartitionByWispSet` (pure-fn, authoritative gate per design bead) | PASS 0.006s | 6 subcases: all_permanent, all_wisps, mixed, empty_input, nil_wisp_set_treats_all_as_permanent, explicit-id_wisp_routes_as_wisp. |
| `TestWispIDSetInTx_HardGate` + `TestWispIDSetInTx_Empty` (Dolt testcontainer, `TESTCONTAINERS_RYUK_DISABLED=true`) | PASS 2.94s | Container flakiness builder reported did not reproduce on deployer pass; matches reviewer's observation. |
| `go test ./internal/storage/issueops/... ./internal/types/... ./internal/ui/...` | PASS | All container-free packages clean. |
## Findings tracked from review
**F1 (medium, scope-completeness — advisory):** Reviewer identified three
additional N-per-ID `IsActiveWispInTx` hot loops the builder did not flag
(`DeleteIssuesInTx`, `GetCommentCountsInTx`, `GetCommentsForIssuesInTx`) on
top of the two already flagged (`GetDependencyRecordsForIssuesInTx`,
`GetBlockingInfoForIssuesInTx`). Design bead be-nu4.2.1 scoped work to the
two named sites, so the build is correct as-delivered; a consolidated
follow-up bead covering all five remaining sites must exist before the
be-nu4 epic closes. Not a release blocker.
**F2 (low, benchmark numbers not captured — advisory):** Benchmark code
ships and compiles; runtime 1K/10K/50K numbers were not captured because
the Dolt server / testcontainer environment is flaky in both the builder
and reviewer sandboxes (Docker 29.4 + ryuk 0.13.0). Release-gate decision:
defer numeric capture to a clean CI environment; code-level deliverable
(the benchmarks exist) is satisfied per ADR §11.2.
## Verdict
**PASS** — push to `fork` (origin is locked for quad341; `fork =
quad341/beads`), open PR against `gastownhall/beads:main`.

View file

@ -1,52 +0,0 @@
# Release Gate: be-i9e2 — bd dolt push no-push unit test
**Branch:** `feat/be-3w6-be-0c8-nopush-dolt`
**Tip commit:** `6a8c547fc5ad63c257272656ab09259c1ce20927`
**Deploy bead:** be-wfb5
**Source bead:** be-i9e2
**Review bead:** be-lijh (PASS)
**Date:** 2026-06-07
## Gate Checklist
| # | Criterion | Result | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | **PASS** | be-lijh reviewed `6a8c547fc` on this branch; verdict PASS. Findings: style/security [info], spec [pass×2], tests [pass]. No [high] findings. |
| 2 | Acceptance criteria met | **PASS** | See below |
| 3 | Tests pass | **PASS** | See below |
| 4 | No high-severity findings open | **PASS** | Review notes contain [info] and [nit] only; zero [high] findings |
| 5 | Final branch is clean | **PASS** | `git status` clean on `feat/be-3w6-be-0c8-nopush-dolt`; no uncommitted changes |
| 6 | Branch diverges cleanly from main | **PASS** | `git merge-tree` reports 0 conflicts with `origin/main` |
| 7 | Single feature theme | **PASS** | All 5 commits are about the no-push guard for `bd dolt push/pull`. Removing any one commit from main would leave the others working independently — but they all form one coherent guard + test bundle. |
## Criterion 2: Acceptance Criteria
From be-i9e2:
- [x] **Focused test asserting `bd dolt push` exits 0 with skip message when `no-push:true`** — `TestNoPushSkipsDoltPush` (`cmd/bd/dolt_test.go:1506`) asserts `fake.pushCalled == false` and output contains `"skipping push"` when `BD_NO_PUSH=true`.
- [x] **Does NOT call the store's Push()** — `minimalPushStore.pushCalled` verified false in the same test.
- [x] **Mirrors `TestNoPushDoesNotSkipDoltPull` pattern** — uses same `saveAndRestoreGlobals(t)` / `t.Setenv` / `config.ResetForTesting()` structure.
## Criterion 3: Test Results
```
$ go test ./cmd/bd/ -run "TestNoPushSkipsDoltPush|TestNoPushDoesNotSkipDoltPull" -v -count=1
WARN: Docker not available, skipping Dolt tests
=== RUN TestNoPushSkipsDoltPush
--- PASS: TestNoPushSkipsDoltPush (0.00s)
=== RUN TestNoPushDoesNotSkipDoltPull
--- PASS: TestNoPushDoesNotSkipDoltPull (0.00s)
PASS
ok github.com/steveyegge/beads/cmd/bd 0.104s
```
Full `cmd/bd` suite run: some tests fail due to **pre-existing environment isolation issues** in the GC rig (live beads DB at `/home/jaword/.beads` interferes with tests that expect no beads present). The same failures are confirmed present on `origin/main` directly — not introduced by this branch. The reviewer ran the full dolt suite on a clean environment and confirmed PASS with no regressions.
## Commits on branch (ahead of origin/main)
| SHA | Message |
|-----|---------|
| `6a8c547` | test(dolt): add TestNoPushSkipsDoltPush for push guard coverage (be-i9e2) |
| `78e882286` | fix(no-push): add missing context and storage imports to dolt_test.go |
| `7836510b0` | fix(no-push): remove pull guard from bd dolt pull (be-ve2x6) |
| `da82de490` | fix(no-push): anchor stripDoltPushReferences to prevent 4-space indent false match |
| `56f7848c7` | feat(no-push): guard dolt push/pull and strip push from rendered templates |

View file

@ -1,25 +0,0 @@
# Release gate: be-iprk — bd show --json count-only + iter flags (PR #4010)
- **Bead:** be-iprk (Review: be-ijck6q bd show --json count-only + iter flags, PR #4010)
- **Branch:** `feat/be-ijck6q-show-json-iter` (HEAD: `0a19244ca`)
- **Evaluated:** 2026-05-19 by beads/deployer
## What ships
- `bd show --json` default now emits `dependent_count`/`dependency_count`/`comment_count` (O(1) COUNT queries) instead of full slices
- `--include-dependents`: streams via `IterDependentsWithMetadata`, shallow-copies each item
- `--include-comments`: streams via `IterIssueComments`
- Storage layer: `CountDependents`, `CountDependencies`, `CountIssueComments`, `CountIssues`, `CountEvents` + Dolt/EmbeddedDolt implementations; PG stubs; OTel telemetry wrappers
## Gate criteria
| # | Criterion | Result | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | **PASS** | be-iprk notes: "VERDICT: pass... DECISION: PASS — label needs-deploy." 2 LOW non-blocking findings documented; no HIGH findings. |
| 2 | Acceptance criteria met | **PASS** | Default JSON output uses COUNT(*) queries; `--include-dependents` and `--include-comments` flags stream via Iter; storage layer additions complete across Dolt/EmbeddedDolt/PG stubs |
| 3 | Tests pass | **PASS** | CI 40/40 SUCCESS on PR #4010 |
| 4 | No high-severity review findings open | **PASS** | 0 HIGH findings; 2 LOW findings in be-iprk: (1) shallowDependentsForJSON defined but not called in production path; (2) defer iter.Close() function-scoped accumulation — both non-blocking |
| 5 | Final branch is clean | **PASS** | `git status` clean |
| 6 | Branch diverges cleanly from main | **PASS** | No conflicts against origin/main |
## Verdict: PASS

View file

@ -1,144 +0,0 @@
# Release Gate: be-jewoem + bd-umbf — reference-aware prune + contributor namespace isolation
**PR**: https://github.com/gastownhall/beads/pull/4023
**Branch**: `feat/be-jewoem-be-u2mw2x-reference-aware-prune` (quad341/beads)
**Gate first evaluated**: 2026-05-17 (be-jewoem/be-u2mw2x scope)
**Gate updated**: 2026-05-18 (bd-umbf Children 1-3 added to branch)
**Gate updated**: 2026-05-19 (PR #4023 description refreshed to cover full bd-umbf scope; CI confirmed all green on run 26010362625)
**Gate updated**: 2026-05-20 (B3 RESOLVED — test commits landed; fixed yaml/DB store mismatch in export exclude-owner and fork-routing config write)
**Gate updated**: 2026-06-06 (re-review PASS confirmed at 9cfe6206f; post-review CI/docs commits noted; CI all green on 0a41b581a; merge-request issued)
---
## Criterion 1: Review PASS
| Scope | Reviewer | Verdict | Date |
|-------|----------|---------|------|
| be-jewoem/be-u2mw2x reference-aware prune | beads/reviewer | **PASS** | 2026-05-17 |
| bd-umbf Children 1-3 (be-c7696f, be-3c5a0f, be-bbj) | beads/reviewer | **request-changes** | 2026-05-17 |
| bd-umbf full scope re-review | beads/reviewer | **PASS** | 2026-05-22 |
Source: bead `be-xbwp` notes (be-jewoem/be-u2mw2x scope) — "REVIEW VERDICT: pass"
Source: bead `be-72b53f` notes (bd-umbf scope) — "REVIEW VERDICT: request-changes"
Source: mail gm-wisp-765hi (2026-05-22) — "Re-review of bd-umbf scope complete. All B1-B4 blockers resolved; CI 41/41 green; 4 LOW non-blocking findings. PR #4023 (commit 9cfe6206f) cleared for deployment." Review bead: be-o5m5. Follow-ups be-zkzw and be-36oa are non-blocking.
bd-umbf blocker resolution status (as of 2026-05-19):
- B1 (doc freshness CI fail): **RESOLVED** — commit e7fbf9b37 regen'd CLI docs; CI now PASS
- B2 (ubuntu-latest CI fail): **RESOLVED** — commit 11d232215 fixed build; CI now PASS
- B3 (missing tests): **RESOLVED** — test commits landed 2026-05-20:
- export_exclude_owner_test.go (TestExportExcludeOwner_{flag,config,verbose}) — cherry-picked; fixed real bug: buildOwnerExcludeSet was not reading yaml-only export.* keys from config.yaml
- migrate_personal_test.go (TestMigratePersonal_{noopWhenEmpty,movesIssues,abortOnNoConfirm}) — cherry-picked; movesIssues skip-logic broadened for Dolt-unavailable environments
- fork_detect_embedded_test.go (TestBdInit_ForkAutoContributor{,_Idempotent,_MaintainerFlag,_NonInteractive}) — cherry-picked; fixed real bug: autoConfigureForkContributor was writing routing.*/sync.* to DB instead of config.yaml (yaml-only keys)
- B4 (no transaction on delete path): **RESOLVED** — commit 11d232215 separates copy/delete phases; DeleteIssues batches delete
Post-review commits (after 9cfe6206f, CI/docs only — no behavioral change):
- `72a2a2521` — chore: merge origin/main into branch (keeps branch current)
- `aea2d3ed7` — fix(ci): register TestPruneLargeFixture in check-testing-short allowlist
- `0a41b581a` — fix(docs): restore version-1.0.5 migrate-personal CLI reference
**→ PASS** (re-review confirmed PASS 2026-05-22; post-review commits are CI/docs only)
---
## Criterion 2: Acceptance Criteria
| # | Criterion | Evidence | Result |
|---|-----------|----------|--------|
| AC-1 | `bd prune` skips referenced closed beads by default | `runPurgeOrPrune`: reference check gated on `scope.cmdName == "prune" && !scope.ignoreReferences` | PASS |
| AC-2 | `bd prune --ignore-references` deletes them anyway | Flag registered in `pruneCmd.init()` only; reviewer confirmed | PASS |
| AC-3 | `bd purge --ignore-references` → "unknown flag" | Flag NOT registered on purgeCmd; reviewer confirmed | PASS |
| AC-4 | `referenced_skipped` + `referenced_count` always in prune JSON (0-included) | Reviewer confirmed "0-included"; `referenced_ids_sample` omitted when 0 | PASS |
| AC-5 | Output matches be-zej8mz §4 spec (all 4 paths) | Reviewer verified against spec | PASS |
| AC-6 | Integration test passes | DEFERRED — reviewer filed be-zkzw; PR has bench test (be-0wt833) only | DEFERRED |
| AC-7 | `go test ./...` passes; golangci-lint clean | All CI test shards PASS; Lint PASS | PASS |
Note on AC-6: Reviewer acknowledged integration-test gap and filed be-zkzw as follow-up. Bench test (be-0wt833) is included and PASS.
**→ PASS** (with AC-6 deferred to be-zkzw)
---
## Criterion 3: Tests Pass
**Initial CI run** on `feat/be-jewoem-be-u2mw2x-reference-aware-prune` (run 26001972966) — be-jewoem/be-u2mw2x scope only:
All shards PASS after doc-freshness fix (commit e15c4c464).
**Latest CI run** (run 26924427657, 2026-06-04) — on current HEAD 0a41b581a (includes post-review CI/docs commits):
| Suite | Result |
|-------|--------|
| Build (Embedded Dolt) | PASS |
| Lint | PASS |
| Check build-tag policy | PASS |
| Check cmd/bd pure-Go tests compile | PASS |
| Check doc flags freshness | PASS |
| Check formatting | PASS |
| Check version consistency | PASS |
| Test (Embedded Dolt Cmd 1–20/20) | PASS |
| Test (Embedded Dolt Storage) | PASS |
| Test (storage domain + uow) | PASS |
| Test (ubuntu-latest) | PASS |
| Test (macos-latest) | PASS |
| Test (Windows smoke) | PASS |
| Test Nix Flake | PASS |
| Upgrade smokes (v1.0.0–v1.0.5) | PASS |
| PR mergeStateStatus | CLEAN |
All 51 CI checks SUCCESS.
**→ PASS** (all CI green as of 2026-06-04 on HEAD 0a41b581a)
---
## Criterion 4: No High-Severity Open Findings
Reviewer notes: "Findings: None blocking."
No HIGH findings in review bead be-xbwp.
**→ PASS**
---
## Criterion 5: Final Branch Clean
`git status` clean after docs regen + gate commit.
**→ PASS**
---
## Criterion 6: Branch Diverges Cleanly from main
GitHub `mergeStateStatus`: **CLEAN** (PR #4023, checked 2026-06-06).
Branch merge base with origin/main: `a5e5cd71f` (Merge pull request #4300 from coffeegoddd/db/list). No textual conflicts; GitHub confirms clean merge.
**→ PASS**
---
## Criterion 7: Single Feature Theme
The PR bundles two tightly coupled features:
- **be-jewoem/be-u2mw2x**: `bd prune` reference-aware skip of closed beads cited by open beads
- **bd-umbf**: contributor namespace isolation — fork auto-configure, export owner filter, migrate-personal
These share the contributor/owner lifecycle subsystem: bd-umbf's `autoConfigureForkContributor`, `export.exclude_owners`, and `migrate-personal` are all prerequisite tooling for correct prune behavior in contributor workflows. The reviewer reviewed and approved the combined scope (gm-wisp-765hi). Neither feature is independently useful without the other in the contributor context.
**→ PASS** (reviewer confirmed combined scope coherent; cleared for deployment as a unit)
---
## Summary
| Criterion | Result |
|-----------|--------|
| 1. Review PASS | **PASS** — re-review 2026-05-22 at 9cfe6206f; 4 LOW non-blocking |
| 2. Acceptance criteria | **PASS** (AC-6 deferred to follow-up be-zkzw) |
| 3. Tests pass | **PASS** (51/51 CI green, run 26924427657, 2026-06-04) |
| 4. No HIGH findings | **PASS** |
| 5. Branch clean | **PASS** |
| 6. Clean divergence from main | **PASS** (mergeStateStatus CLEAN) |
| 7. Single feature theme | **PASS** (coupled contributor namespace features) |
**Overall: PASS — re-review confirmed 2026-05-22, all CI green 2026-06-04 on HEAD 0a41b581a. PR #4023 ready for merge. Merge authority: mayor/mpr.**

View file

@ -1,84 +0,0 @@
# Release gate — be-l9q (`bd list --skip-labels` AD-02)
**Date:** 2026-04-27
**Deployer:** beads/deployer (deployer-1, second pass after builder rebase)
**Bead (review):** be-l9q — Review: bd list --skip-labels (AD-02)
**Feature bead:** be-a5z (closed)
**Builder commit (post-rebase):** `5e6d84b4` (was `bbfa50df` pre-rebase)
**Source branch:** `be-vzu-rebase-fix` (builder worktree, rebased onto current `origin/main`)
**Final branch:** `release/be-l9q` @ `c4b7e0ac` (cherry-pick of `5e6d84b4` onto `origin/main`)
**Base:** `origin/main` @ `f4c46d91` ("/go.{mod,sum}: bump dolt driver (#3435)")
## Verdict: PASS
## What this ships
`bd list --skip-labels` — a hydration toggle for the list command that:
- Skips the `GetLabelsForIssuesInTx` SQL JOIN at the search-layer when set,
with a defense-in-depth gate at the `cmd/bd` post-query step too.
- Emits `labels: []` in JSON via a `skipLabelsIssueView` wrapper that
overrides the `omitempty` on `Issue.Labels`, so machine consumers can
always parse the field.
- Surfaces the suppression in human output: footer in pretty/tree and
buffered/compact paths (gated by `!isQuiet()`); `--long` swaps the
Labels block to `Labels: (suppressed by --skip-labels)`.
- Refuses combination with the six existing label-filter flags with
exit code 2 and a Wireframe-5 error message.
## Criteria
| # | Criterion | Result | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | PASS | reviewer-1 PASS + reviewer-2 (concur) PASS in be-l9q notes. |
| 2 | Acceptance criteria met | PASS | All 5 ACs walked + Wireframes 1-6 covered in reviewer-1's matrix. |
| 3 | Tests pass | PASS | Targeted suite all green; broader sweep failures are pre-existing on `origin/main` (see below). |
| 4 | No HIGH-severity review findings open | PASS | No HIGH findings — one LOW advisory (storage-gate test is behavioral not query-counter) is non-blocking. |
| 5 | Final branch is clean | PASS | `git status` clean after gate commit; only untracked items unrelated to this bead. |
| 6 | Branch diverges cleanly from main | PASS | `git cherry-pick 5e6d84b4` onto `origin/main` applied without conflict. All 7 touched files were byte-identical on `origin/main` and `5e6d84b4^`, so the rebased commit lands without drift. |
## Test evidence (criterion 3)
Run from `release/be-l9q` @ `c4b7e0ac`:
- `go vet -tags gms_pure_go ./cmd/bd/... ./internal/storage/issueops/... ./internal/types/...`: clean (exit 0).
- Targeted: `go test -tags gms_pure_go -run 'TestFormatIssueLong|TestSkipLabels|TestFormatSkipLabelsConflictError' ./cmd/bd/`: PASS (1.081s). Includes:
- `TestSkipLabelsConflicts` (8 subs)
- `TestSkipLabelsIssueView_AlwaysEmitsLabelsArray`
- `TestFormatSkipLabelsConflictError`
- `TestFormatIssueLong*` (open/closed/with assignee/with labels/with metadata variants)
- Broader sweep: `go test -tags gms_pure_go -count=1 ./cmd/bd/ ./internal/storage/issueops/ ./internal/types/`:
- `internal/storage/issueops`: PASS
- `internal/types`: PASS
- `cmd/bd`: 2 failures, both **pre-existing on `origin/main`**:
- `TestWhereCommand_ReadsPrefixFromEmbeddedStore` — verified FAIL on bare `origin/main` (`f4c46d91`) checkout.
- `TestResolveWhereBeadsDir_UsesInitializedDBPath` — passes in isolation; fails only when run alongside the above due to shared test state. Same behavior on bare `origin/main`.
- Both failures are in `bd where` test infrastructure and have no code path through `cmd/bd/list.go`, `internal/storage/issueops/search.go`, or any file touched by `5e6d84b4`. Builder and reviewer-2 both flagged this `cmd/bd` test-suite environmental flakiness on parent commits in their notes.
## Cherry-pick mechanics (criterion 6)
The first deployer pass FAILed criterion 6 because the original `bbfa50df`
conflicted with upstream `bb6b8f22` (#3481, "Fix ready/list UX regressions")
in `cmd/bd/list_format.go`. The builder rebased `be-vzu-rebase-fix` onto
current `origin/main` (`f4c46d91`), resolved the conflict in the
`formatIssueLong` Description-block-vs-`labelsSkipped` interleaving (per the
suggested form in the prior FAIL note), and produced `5e6d84b4` as the new
be-a5z commit.
The deployer cherry-picked just `5e6d84b4` (not the entire 6-commit branch)
because the review of be-l9q covers only the be-a5z change. Pre-flight check
confirmed every file touched by `5e6d84b4` is byte-identical between
`origin/main` and `5e6d84b4^` (the rebased be-a5z parent in the builder
chain), so cherry-picking the single commit produces the same tree it would
on the rebased branch. Cherry-pick applied without conflict; `git status`
clean.
| File | `origin/main` blob | `5e6d84b4^` blob |
|---|---|---|
| `cmd/bd/list.go` | `54c277a1` | `54c277a1` |
| `cmd/bd/list_format.go` | `e723a973` | `e723a973` |
| `cmd/bd/list_skip_labels_test.go` | (new) | (new) |
| `cmd/bd/list_test.go` | `e4640f29` | `e4640f29` |
| `cmd/bd/show_format_metadata_test.go` | `cc988f92` | `cc988f92` |
| `internal/storage/issueops/search.go` | `b9d297ca` | `b9d297ca` |
| `internal/types/types.go` | `45f902bf` | `45f902bf` |

View file

@ -1,49 +0,0 @@
# Release gate: be-ow4w54 — Extract markdown rendering out of internal/ui
**Verdict: PASS for the internal/ui decoupling scope.**
Scope note: this gate verifies that `internal/ui` no longer imports the
glamour/chroma markdown renderer stack and that direct markdown-rendering
callers use `internal/uimd`. It does not prove that every `bd` subprocess avoids
glamour/chroma startup work; `cmd/bd` is a single Go binary, so that broader
runtime claim needs a separate startup/dependency boundary and measurement.
Branch: `feat/be-ow4w54-uimd-markdown`
Base on origin: `origin/main` (merge-base `74c66722b14cc32e70ecc4182788d4de40b6d906`)
Reviewed contributor head: `228ead67e34d2ce77254d65a5c3bdb98b7854bbf`
Source bead: `be-ow4w54`; review bead: `be-2y5h4f` (PASS).
## Commits
| # | SHA | Subject |
|---|-----|---------|
| 1 | `228ead67e` | refactor(ui): extract markdown rendering to internal/uimd package |
| 2 | local maintainer fixup | docs: narrow uimd release gate scope |
Original contributor change: `cmd/bd/comments.go`, `cmd/bd/show.go`,
`cmd/bd/show_children.go`, `cmd/bd/show_display.go`,
`internal/{ui→uimd}/markdown.go`.
Maintainer fixup: narrows this gate to the delivered `internal/ui`
decoupling scope and restores rationale comments in
`internal/uimd/markdown.go`.
## Criteria
| # | Criterion | Verdict | Evidence |
|---|-----------|---------|----------|
| 1 | Review PASS present | **PASS** | be-2y5h4f notes: "VERDICT: pass. Findings: none." |
| 2 | Acceptance criteria met | **PASS** | `internal/ui` contains no glamour/chroma imports (`grep -rn 'glamour\|chroma' internal/ui/` -> empty). `internal/uimd/markdown.go` created and all 4 direct markdown callers (show.go, show_display.go, show_children.go, comments.go) updated to import `internal/uimd`. Build clean. |
| 3 | `internal/ui` dependency graph stays renderer-free | **PASS** | `go list -deps ./internal/ui \| grep -Ec 'charm\.land/glamour\|alecthomas/chroma'` -> `0`. |
| 4 | Tests pass | **PASS** | `go test -tags gms_pure_go -count=1 -short ./internal/ui/... ./internal/uimd/...`: ok (0.005s). `go test -tags gms_pure_go -count=1 -short -run TestShow ./cmd/bd/`: ok (0.335s). `go build -tags gms_pure_go ./...`: clean. Note: `TestCheckBeadsRole_*` failures in `cmd/bd/doctor` are pre-existing rig-env failures unrelated to this change. |
| 5 | No high-severity review findings open | **PASS** | Reviewer be-2y5h4f: "Findings: none." |
| 6 | Final branch is clean | **PASS** | `git status` clean (untracked `.gc/`, `.gitkeep` are rig scaffolding). |
| 7 | Branch diverges cleanly from main | **PASS** | Contributor head rebased cleanly onto `origin/main`; maintainer fixup remains local for the next review pass. |
## Test environment
- Host: Linux 6.19.14-300.fc44.x86_64
## Push target
`PUSH_REMOTE=fork`. Cross-repo PR head: `quad341:feat/be-ow4w54-uimd-markdown`.

View file

@ -1,49 +0,0 @@
# Release gate: be-psutr9 — cobra/pflag race-fix (cherry-pick from Julian Knutsen)
**Verdict: PASS.**
Branch: `fix/cobra-pflag-race-d8a97c6bb`
Base branch: `main` (rebased review worktree merge-base: `39f3148fc`)
Reviewed worktree head before workflow fixups: `cd504231d`
Original PR head: `d8c7cc56c`
Review bead: `be-psutr9` (initial PASS); re-review: `be-mouz` (PASS, 2026-05-19).
## Commits
| # | SHA | Subject |
|---|-----|---------|
| 1 | `99ee12d29` | test: prevent Cobra command flag cache races |
| 2 | `06e8d98d2` | chore: release gate PASS for be-psutr9 (cobra/pflag race-fix) |
| 3 | `cd504231d` | docs(gate): update be-psutr9 gate — re-review PASS (be-mouz), fix stale file list |
Original upstream commit: `d8a97c6bb` by Julian Knutsen. The rebased code
commit in this worktree is `99ee12d29`; the gate commits are maintainer
release documentation layered on top of the contributor code change.
Files changed in the review range:
`cmd/bd/stdio_race_guard_test.go`, `cmd/bd/test_helpers_pure_test.go`, and
`release-gates/be-psutr9-gate.md`.
**Note:** Runtime deparallelization for `TestStaleCommandInit` and
`TestNotionCommandsRegistered` had already landed on `origin/main` before this
PR's merge base. This PR is the static guard extension that prevents future
parallel tests from reintroducing the Cobra flag-cache race.
## Criteria
| # | Criterion | Verdict | Evidence |
|---|-----------|---------|----------|
| 1 | Review PASS present | **PASS** | be-psutr9 initial PASS (no findings); re-review be-mouz PASS (2026-05-19): "Correct fix for pflag lazy-merge race. No production code changes." LOW: stale gate file resolved in this update. |
| 2 | Acceptance criteria met | **PASS** | (a) The runtime deparallelization was already present on `origin/main`; this PR extends the static guard for the root-cause Cobra methods. (b) `cobraOutputMethods` renamed to `cobraParallelUnsafeMethods`; `.Find(` and `.InheritedFlags(` added so the policy guard catches future parallel callers. (c) Test (macos-latest) with race detector PASS in CI (6m2s), confirming the flake is gone. |
| 3 | Tests pass | **PASS** | CI on PR #4011: Test (macos-latest) PASS 6m2s, Test (ubuntu-latest) PASS 10m24s, Test (Windows-smoke) PASS 3m1s, Test Nix Flake PASS 3m35s, Lint PASS, all other checks PASS. Test (Embedded Dolt Cmd 17/20) FAIL — assessed as known `bd-embedded-test-json-stderr-leak` flake; PR touches no embedded test helper. |
| 4 | No high-severity review findings open | **PASS** | Reviewer be-psutr9: "Findings: none." No security surface — test-only change. |
| 5 | Final branch is clean | **PASS** | Review worktree status was clean except workflow-local `.gc/` review artifacts. |
| 6 | Branch diverges cleanly from main | **PASS** | Review range from `39f3148fc..cd504231d` contained the Cobra guard commit plus gate documentation only. `git merge-tree` showed zero conflicts. |
## Push target
`PUSH_REMOTE=fork`. PR #4011 opened within fork: `quad341:fix/cobra-pflag-race-d8a97c6bb` → `gastownhall/beads:main`.
## Flake note
The single CI failure (Test Embedded Dolt Cmd 17/20) is the known `bd-embedded-test-json-stderr-leak` flake — a test helper uses `cmd.CombinedOutput()` causing auto-export warnings on stderr to leak into JSON stdout. This PR touches none of the relevant files (`cmd/bd/*_embedded_test.go` helper). See bd memory `bd-embedded-test-json-stderr-leak`.

View file

@ -1,88 +0,0 @@
# Release gate — be-ripy (be-avn `staleDatabasePrefixes` canonical sync)
**Date:** 2026-04-30
**Deployer:** beads/deployer (deployer-1, fresh evaluation post-builder-rebase)
**Bead (review):** be-ripy — Review: be-avn staleDatabasePrefixes canonical sync (AD-04 follow-up)
**Feature bead:** be-avn (closed; source bead)
**Reviewed commit:** `0f76453f` (rebased equivalent: `ee67d5f9` on `be-vzu-rebase-fix`)
**Final branch:** `release/be-ripy` @ `1014c759` (cherry-pick of `ee67d5f9` onto `origin/main`)
**Base:** `origin/main` @ `8694c535` ("doctor: detect AGENTS.md / CLAUDE.md user-authored divergence (#3600)")
## Verdict: PASS
## What this ships
`bd dolt clean-databases` cleanup-side prefix list converged with the
firewall-side `testDatabasePrefixes`. After this change:
- `staleDatabasePrefixes` becomes `[testdb_, beads_test, beads_pt, beads_vr,
doctest_, doctortest_, benchdb_]` — the architect-canonical 7-prefix list.
- Net behavioral change vs. prior list:
- **+** `benchdb_` now cleaned (closes the AD-04-documented bench-DB leaker)
- **+** `beads_test` now cleaned (matches firewall scope when be-c5p lands)
- **−** `beads_t<hex>` protocol-test prefix no longer matched by cleanup;
those workspaces clean themselves up in tearDown per AD-04 problem statement
- The doc-block on `staleDatabasePrefixes` is restructured to mirror
`testDatabasePrefixes` (per-prefix origin notes plus a cross-reference
to the two sibling lists). The `Long:` help string is updated.
## Criteria
| # | Criterion | Result | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | PASS | reviewer-gm-po6pox3 PASS verdict in be-ripy notes — 0 blockers, 0 high, 0 medium, 1 advisory low (out-of-scope doctor copy follow-up). |
| 2 | Acceptance criteria met | PASS | All 4 ACs walked by reviewer (canonical list match, sibling list already matches, cross-reference doc-block, build/lint clean) — re-verified below. |
| 3 | Tests pass | PASS | Build + vet clean on full repo; targeted tests PASS (see below). |
| 4 | No HIGH-severity review findings open | PASS | 0 HIGH findings. The 1 LOW advisory (doctor's separate `staleDatabasePrefixes` is not converged) is explicitly out-of-scope per the bead's "Files in scope". |
| 5 | Final branch is clean | PASS | `git status` clean on `release/be-ripy` (only untracked items are `.gc/` and `.gitkeep` from worktree harness, unrelated to this change). |
| 6 | Branch diverges cleanly from main | PASS | `git cherry-pick ee67d5f9` onto `origin/main@8694c535` applied without conflict. Only one file changed (`cmd/bd/dolt.go`); no path overlap with the 32 commits on `origin/main` since the prior rebase base `f4c46d91`. |
## Test evidence (criterion 3)
Run from `release/be-ripy` @ `1014c759`:
- `go build -tags gms_pure_go ./...` — clean (exit 0).
- `go vet -tags gms_pure_go ./...` — clean (exit 0).
- `go test -tags gms_pure_go -count=1 -run 'TestStaleCommand|TestDoltClean' ./cmd/bd/` — PASS (0.089s). Includes `TestStaleCommandInit` (Docker-gated `TestStaleSuite` skipped — Docker not available in deployer harness; runs on CI).
- `go test -tags gms_pure_go -count=1 -run 'TestStaleDatabasePrefixes' ./cmd/bd/doctor/` — PASS (0.073s, 12 sub-tests). The doctor package has its own `staleDatabasePrefixes` list (out-of-scope per AD-04 bead), confirmed unaffected.
The reviewer's targeted suite (`go build`, `golangci-lint run`) was independently re-verified at the source SHA before rebase. No new code paths in this commit — the change is a constant-list edit plus a doc-block rewrite, so existing coverage applies.
## Cherry-pick mechanics (criterion 6)
This deploy follows the precedent of PR #3562 / `release/be-l9q`: the bead
covers a single commit, so the deployer cherry-picks just that commit onto
fresh `origin/main` rather than shipping the full source branch. The
reviewed SHA `0f76453f` was rebased by the builder to `ee67d5f9` after the
gate-FAIL on the prior `be-l9q` deploy attempt; the builder confirmed
content equivalence (`git diff ee67d5f9 0f76453f` shows only an unrelated
`.gitattributes` upstream-noise hunk).
Cherry-pick on top of current `origin/main` (8694c535) was conflict-free.
Touched file (`cmd/bd/dolt.go`) was not modified by any of the 32 commits
on `origin/main` since the prior rebase base, so the rebased commit's tree
applies textually.
## Cross-list convergence note (informational)
The new doc-block declares the cleanup list, the firewall list
(`internal/storage/dolt/store.go:testDatabasePrefixes`), and the city
formula list "must converge". On current `origin/main`, the firewall
list does not yet contain `benchdb_` (that addition lands with be-c5p,
not yet shipped). After this deploy:
- Cleanup list (this PR): `[testdb_, beads_test, beads_pt, beads_vr,
doctest_, doctortest_, benchdb_]` — 7 prefixes
- Firewall list (`origin/main`): `[testdb_, beads_test, beads_pt,
beads_vr, doctest_, doctortest_]` — 6 prefixes (missing `benchdb_`)
The two lists drift on `benchdb_` until be-c5p ships. This is a
defense-in-depth gap (firewall doesn't reject `benchdb_*` at creation
time on the production server), not a runtime regression — the cleanup
job still drops them. Not a gate blocker.
## Hand-off
- Final branch `release/be-ripy` @ `1014c759` ready to push.
- PR target: `gastownhall/beads:main` (origin), pushed via `quad341/beads`
(fork) per the rig's push-target protocol.

View file

@ -1,39 +0,0 @@
# Release Gate: be-wqt8 — fork auto-config planning store init fix
**Bead:** be-chhm (deploy bead for be-mx1y / be-wqt8)
**Branch:** fix/be-wqt8-planning-init
**Cherry-picked from:** feat/be-3w6-be-0c8-nopush-dolt
**Commit:** 3cb77b014 (cherry-pick of 1d9c835a2)
**Date:** 2026-06-02
## Gate Result: PASS
| # | Criterion | Status | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | PASS | be-mx1y: reviewer PASS verdict in bead notes. All 3 acceptance criteria verified by reviewer. |
| 2 | Acceptance criteria met | PASS | See below |
| 3 | Tests pass | PASS | `make test` exit 0 on fresh branch off origin/main. Coverage 37.3%. |
| 4 | No high-severity review findings open | PASS | Reviewer posted 5 findings, all PASS. Zero HIGH or MEDIUM severity findings. |
| 5 | Final branch is clean | PASS | `git status` clean, working tree clean, nothing to commit. |
| 6 | Branch diverges cleanly from main | PASS | Cherry-picked cleanly onto origin/main (fbcee6c62). No conflicts. |
| 7 | Single feature theme | PASS | Single commit touching `cmd/bd/init_contributor.go` and `cmd/bd/init_embedded_test.go`. One subsystem (fork auto-config init path). |
## Acceptance Criteria
1. **Planning Dolt schema initialized during fork auto-config** → SATISFIED
`init_contributor.go:331-337`: `newDoltStoreFromConfig` called on `planningBeadsDir` immediately after creation. CGO builds initialize schema; no-CGO builds skip silently (non-fatal).
2. **YAML-configured contributor setup covered by idempotency guard** → SATISFIED
`init_contributor.go:295-303`: new check on `config.GetValueSource("routing.contributor")` catches `SourceConfigFile` and `SourceEnvVar` — correctly broader than DB-only check.
3. **Regression test exercises the fix** → SATISFIED
`init_embedded_test.go:369-373`: asserts `embeddeddolt` dir exists after `bd init` on fork, proving `initSchema` ran. Gated `//go:build cgo`. No-CGO builds compile clean.
## Test Notes
- `make test` passes clean on a fresh worktree off `origin/main` with this commit cherry-picked.
- The builder worktree's `../../bd` binary (v1.0.4, no-CGO, built May 20) caused unrelated test failures when tests ran there — these are NOT regressions from this commit. Verified by running the same tests in a temp worktree at both `31e416872` (parent) and `1d9c835a2` (this commit) without the stale binary, both pass.
## Deploy Notes
The source branch `feat/be-3w6-be-0c8-nopush-dolt` fork remote has diverged from the local branch (non-fast-forward; builder rebase not pushed to fork). The commit deploys cleanly as a standalone cherry-pick because it is independent of the no-push feature in that branch. PR #4212 continues to track the no-push feature separately.

View file

@ -1,44 +0,0 @@
# Release Gate: be-xra8 + be-roho — bd preflight --fix mode
**PR**: https://github.com/gastownhall/beads/pull/4054
**Branch**: `feat/be-xra8-be-roho-preflight-fix-mode` (quad341/beads)
**Date**: 2026-05-21
**Deployer**: beads/deployer
## Gate Result: PASS
| # | Criterion | Evidence | Result |
|---|-----------|----------|--------|
| 1 | Review PASS present | be-nebh: PASS — "CI 41/41; no injection vectors; sentinel-hash trick correct; deferred restore handles all early-exit paths" — 2 LOW notes, no blockers | ✅ PASS |
| 2 | Acceptance criteria met | See below | ✅ PASS |
| 3 | Tests pass | CI run 26188217144 — all 41 checks PASS | ✅ PASS |
| 4 | No high-severity findings | be-nebh: no blockers — LOW-1: fixVersionSync Skipped Detail empty when default.nix absent; LOW-2: fixNixHash fallback regex could pick wrong hash if nix format changes | ✅ PASS |
| 5 | Final branch is clean | `git status` — clean; only rig artifacts untracked | ✅ PASS |
| 6 | Branch diverges cleanly from main | 3 commits ahead of `origin/main`; merge state CLEAN | ✅ PASS |
## Acceptance Criteria Verification (be-xra8 + be-roho)
| Criterion | Evidence | Result |
|-----------|----------|--------|
| `fixNixHash()` implemented — sentinel-hash trick | Function at line 634 in `cmd/bd/preflight.go`; writes `sha256-AAA...` sentinel, runs `nix build .#default`, parses `got: sha256-...` from error output | ✅ |
| `fixNixHash()` skips when `go.sum` has no uncommitted changes | Checks both `git diff HEAD -- go.sum` and `git diff --cached -- go.sum` before proceeding | ✅ |
| `fixNixHash()` exits with manual instructions when `nix` not in PATH | `exec.LookPath("nix")` guard | ✅ |
| `fixNixHash()` preserves original file permissions | `os.Stat` before write; restores with `os.Chmod` | ✅ |
| `fixVersionSync()` implemented — reads version from `default.nix`, updates `version.go` | `func fixVersionSync()` at line in `cmd/bd/preflight.go` | ✅ |
| `fixVersionSync()` skips gracefully when `default.nix` absent | `os.Stat(vgoPath)` check; returns nil (not error) when absent | ✅ |
| `--fix` flag registered on `bd preflight` | `func runFixes(jsonOutput bool)` at line 572; wired via `--fix` flag | ✅ |
| CLI docs regenerated (be-laor blocker resolved) | `5459e436f docs(preflight): regenerate website CLI reference for --fix flag` + `08c428165 docs: sync CLI_REFERENCE.md --fix help text` | ✅ |
| No user-controlled path injection | Reviewer confirmed: `exec.Command` args are fixed strings; `planningPath` is directory path only | ✅ |
**Note on LOW-1** (reviewer): `fixVersionSync` Skipped Detail empty when `default.nix` absent — the `Skipped` status is set but `Detail` is not populated. Cosmetic; non-blocking.
**Note on LOW-2** (reviewer): `fixNixHash` fallback regex `got: sha256-...` could pick wrong hash if nix output format changes. Acceptable risk — the regex matches standard nix hash-mismatch error format documented in `pr-audit-nix-vendorhash-drift-pattern` memory.
**Note on tests**: be-b6m9 (tests for `--fix` mode) is in progress with validator on a separate branch; follow-up PR expected. Reviewer accepted this.
## Commits
| SHA | Description |
|-----|-------------|
| `0f88c3828` | feat(preflight): implement --fix mode for vendorHash and version sync |
| `08c428165` | docs: sync CLI_REFERENCE.md --fix help text with live CLI (be-xra8, be-roho) |
| `5459e436f` | docs(preflight): regenerate website CLI reference for --fix flag (be-laor) |

View file

@ -1,19 +0,0 @@
# Release gate: be-z33z — CONTRIBUTING_PR_GUIDELINES.md (PR #3913)
- **Bead:** be-z33z (Review: CONTRIBUTING_PR_GUIDELINES.md, PR #3913)
- **Branch:** `docs/be-ad6qbk-contributing-pr-guidelines`
- **Head commit:** `2bbcc1f9d` — docs: add CONTRIBUTING_PR_GUIDELINES.md (codify reviewer expectations for contributors)
- **Evaluated:** 2026-05-19 by beads/deployer
## Gate criteria
| # | Criterion | Result | Evidence |
|---|-----------|--------|----------|
| 1 | Review PASS present | **PASS** | be-z33z notes: "Review verdict: PASS. Docs-only PR... Content is accurate and matches observed reviewer expectations. No security surface, no code change. Routing to deployer." |
| 2 | Acceptance criteria met | **PASS** | `docs/CONTRIBUTING_PR_GUIDELINES.md` added (39 lines); `PR_MAINTAINER_GUIDELINES.md` cross-reference valid |
| 3 | Tests pass | **PASS** | CI 18/18 non-skipped SUCCESS, 3 SKIPPED (Embedded Dolt build/test jobs expected skip for docs-only PR) |
| 4 | No high-severity review findings open | **PASS** | 0 findings in be-z33z |
| 5 | Final branch is clean | **PASS** | `git status` clean |
| 6 | Branch diverges cleanly from main | **PASS** | 1 commit ahead of base (no conflicts) |
## Verdict: PASS