design-playbook 0.14.0 → 0.20.0

package/commands/design-io.md

@@ -1,5 +1,5 @@
 ---
-description: Run Design I/O end-to-end (reference-intake? → spec? → plan? → shell → preview* → fill → craft → accept)
+description: Run Design I/O end-to-end with run-profile tiering (P1/P2/P3) and the recirculate loop (reference-intake? → spec?/shaping? → plan → shell → preview* → fill → craft → accept)
 ---
 
 Run skill **design-playbook** in full. Honor each step’s completion criterion before the next. Recirculate blocking evaluator findings to the owning declaration. Entry routing and plan/preview orchestration live in that skill (not here).

package/commands/run-review.md

@@ -1,5 +1,5 @@
 ---
-description: Cross-run review of Design I/O runs — gate table + repeat blockers
+description: Cross-run review of Design I/O runs — gate table + repeat blockers + rule candidate queue (derived, report-only)
 ---
 
 Cross-run **run review** over `.scratch/<run>/` dirs in the user project. Not a step of a single Design I/O run. Markdown only; report header **`run-review/v1`**.
@@ -13,8 +13,9 @@ Scan user-side `.scratch/<run>/` (not monorepo `dogfood/*` globs). **Include** o
 1. **Inclusion manifest** first: `path | status` (`included` / `skipped` + reason). Note: hash match ≠ honest transcription of `observed`.
 2. **Per-run table** — mandatory **run-path** column; other columns as needed; **gate** from real `validate_run.py` exit when the script is present (plugin install: `packages/design-playbook/scripts/validate_run.py` per run); else literal `not checked`. **Never** infer ok from "artifacts look complete".
 3. **Repeat blockers** — pure frequency table `count | runs | observed text` (verbatim first-seen text). A **repeat blocker** is the same normalized `observed` text recurring across runs (**counting, not judging**). Rows only where ledger `result != pass`. Grouping key = `observed` **casefold + whitespace-collapsed**, then **char-for-char** equality only; `count ≥ 2`. Literal differences stay separate; optional `similar:` pointer line, never merge counts. **`_none_` is normal** when nothing qualifies (do not loosen normalization to manufacture repeats).
-4. **Point-back** cites: path + verbatim `observed:` quote; no line numbers.
-5. Rollup numbers derived **row-by-row** from the tables above — no "overall it seems".
+4. **Rule candidate queue** (derived view, protocol — vNext S5) — derived from the point-back **findings** (not the ledger): group findings by normalized `issue` text (same normalization as repeat blockers), then a candidate enters the queue when **distinct runs ≥ 3 AND distinct task contexts ≥ 2 AND unexplained false positives = 0**. Task context per occurrence comes from the run's contract / spec / manifest method-semantics keys (user / task / environment / method) — **repeats with different contexts are never merged**; a corpus without readable contexts reports the context gap instead of qualifying. Show qualifying candidates (`candidate id | runs | contexts | occurrences`), plus below-threshold signals with their gap list (e.g. `distinct_runs 2 < 3`) — the queue reports distance to qualification, never silently drops it. **Report only**: candidates are never written back to the registry, the governance log, or the baseline; promotion is a user decision recorded in `<project>/rules-governance.jsonl` (append-only; agent may not write adjudication events).
+5. **Point-back** cites: path + verbatim `observed:` quote; no line numbers.
+6. Rollup numbers derived **row-by-row** from the tables above — no "overall it seems".
 
 Ledger row shape: **ui-evaluator** step 2 (do not restate).
 
@@ -24,6 +25,7 @@ Ledger row shape: **ui-evaluator** step 2 (do not restate).
 - no prose lessons / narrative "learning"
 - no auto-writeback to baseline
 - no semantic clustering of observed — verbatim grouping only
+- no auto-promotion — candidates are reported, never written to the registry or governance log
 ```
 
 Scope:

package/commands/ui-review.md

@@ -1,5 +1,5 @@
 ---
-description: Declaration-backed UI review with point-back findings
+description: Dual-track UI review emitting the six-block point-back report (ledger / findings / positives / coverage / limitations / verdict)
 ---
 
 Run skill **ui-evaluator** (pull craft-guard checks when AI slop/motion/loading is in scope). Output issue/source/fix/severity; blocking first.

package/commands/ux-spec.md

@@ -1,5 +1,5 @@
 ---
-description: Six-layer spec.md only (stop before UI shell/code)
+description: Six-layer spec.md via the S0-S6 shaping session (question/assumption/confirmation batches + session artifacts); stop before UI shell/code
 ---
 
 Run skill **ux-spec** only. Emit complete `spec.md`. Do not pick templates or write UI.

package/examples/README.md

@@ -5,6 +5,10 @@ Self-authored samples for onboarding and dogfood. **Not** ports of any upstream
 | File | Use |
 | --- | --- |
 | [ops-list-spec.md](ops-list-spec.md) | Minimal six-layer spec (ops delivery log) |
+| [export-entry/](export-entry/) | P2 fixture run walkthrough: shaping session (S0-S6) + registry craft audit + six-block point-back with one Recirculate round; decision report carries R/C-tier DD entries (S2) |
+| [export-upgrade/](export-upgrade/) | P3 fixture run walkthrough (S2; obligations completed in S6): E-tier DD entries with two preview confirmation rounds, provider candidate asset reference, R3 `dd:` challenge with supersedes revision + invalidated set, and baseline-drift stale review (keep exit); full-catalog craft audit + 10/10 sampling matrix |
+| [dogfood/](dogfood/) | S6 dogfood fixture: a full P3 chain over this repo's own showcase queue surface (run-profile P3, shaping projection, E-tier DD with two preview rounds, R3 re-entry + one R4 repair, method-semantics manifest, full-catalog craft audit, 10/10 sampling matrix, verdict Pass) — static synthesized, no chromium |
+| [rules-governance/](rules-governance/) | Paper governance-log walkthrough (S5): candidate opened from a 3-run / 2-context / 0-false-positive signal, then user promote -> advisory -> machine-enforced with the six criteria records |
 | [settings-decision-report.md](settings-decision-report.md) | ui-picker decision report before code |
 | [point-back-findings.md](point-back-findings.md) | evaluator output shape + recirculate |
 

package/examples/craft-detectors/composition-contrast.md

@@ -1,16 +1,18 @@
-# Composition detector contrast cases
+# Composition craft contrast cases
 
-Authored worked cases for CRAFT-01 through CRAFT-05. Each case combines a rendered observation with independent source evidence. `Expected` is advisory detector status, not evaluator severity or verdict.
+Authored worked cases for CRAFT-01 through CRAFT-05 against the registry. Each applicable case combines a rendered observation with independent source evidence. `Result` is the advisory audit outcome, not evaluator severity or verdict.
 
-| Case | ID | Expected | Rendered evidence | Source evidence | Exception check | Owner hint | Positive fix |
-| --- | --- | --- | --- | --- | --- | --- | --- |
-| competing-actions-hit | CRAFT-01 | hit | Header exposes Create, Import, and Export with equal filled emphasis | All three actions use primary Button variant | Spec names Create as main task; no equal-choice exception | craft | Keep Create primary and group Import/Export as secondary actions |
-| equal-choice-clear | CRAFT-01 | clear | Two plan options receive equal weight in a deliberate comparison | Shared selectable-card primitive applies same selected affordance | Spec requires neutral side-by-side choice before Continue | craft | - |
-| queue-card-wall-hit | CRAFT-02 | hit | Twelve equal cards make status and owner hard to compare | Queue records map directly to Card wrappers | Records are operational rows, not independent browse objects | template | Replace card wall with table/list and stable comparison columns |
-| card-collection-clear | CRAFT-02 | clear | Gallery cards represent independent templates with image, owner, and open action | TemplateCard owns preview and item-level interaction | Browsable independent objects fit card semantics | template | - |
-| nested-panels-hit | CRAFT-03 | hit | Page section floats as a shadowed card containing three more bordered cards | Card wraps section and nested Cards provide spacing only | No modal, tool, or repeated-item semantics justify frames | template | Flatten section, group with spacing/dividers, retain frame only for tool |
-| drawer-tool-clear | CRAFT-03 | clear | Unframed page contains one bounded detail drawer | Drawer primitive owns focus and dismiss behavior; no nested Card | Detail drawer is a genuine framed tool | components | - |
-| blue-everywhere-hit | CRAFT-04 | hit | Canvas, surfaces, actions, charts, and state badges use blue variants | Single blue ramp fills surface, action, success, and selected roles | No verified monochrome baseline or semantic role separation | design | Restore neutral surfaces and assign distinct semantic state roles |
-| semantic-color-clear | CRAFT-04 | clear | Neutral surfaces dominate; cyan marks selection, amber warning, red failure | Tokens map each semantic role independently | Brand accent remains selective and states stay distinguishable | design | - |
-| pill-everything-hit | CRAFT-05 | hit | Tabs, buttons, fields, filters, and section labels all use pill silhouettes | Global radius-full token applies across unrelated controls | No baseline or semantic grouping justifies uniform pill geometry | components | Restore control-specific shapes and reserve pills for compact tags |
-| status-pill-clear | CRAFT-05 | clear | Compact pills appear only for status and removable filter chips | Badge and Chip primitives use pill shape; buttons use restrained radius | Shape communicates status/grouping and is not page-wide | components | - |
+| Case | ID@ver | Applicability | Predicate reason / missing proof | Result | Rendered evidence | Source evidence | Exception check | Positive fix |
+| --- | --- | --- | --- | --- | --- | --- | --- | --- |
+| competing-actions-hit | CRAFT-01@1 | applicable | - | hit | Header exposes Create, Import, and Export with equal filled emphasis | All three actions use primary Button variant | Spec names Create as main task; no equal-choice exception | Keep Create primary and group Import/Export as secondary actions |
+| equal-choice-clear | CRAFT-01@1 | applicable | - | clear | Two plan options receive equal weight in a deliberate comparison | Shared selectable-card primitive applies same selected affordance | Spec requires neutral side-by-side choice before Continue | - |
+| queue-card-wall-hit | CRAFT-02@1 | applicable | - | hit | Twelve equal cards make status and owner hard to compare | Queue records map directly to Card wrappers | Records are operational rows, not independent browse objects | Replace card wall with table/list and stable comparison columns |
+| card-collection-clear | CRAFT-02@1 | applicable | - | clear | Gallery cards represent independent templates with image, owner, and open action | TemplateCard owns preview and item-level interaction | Browsable independent objects fit card semantics | - |
+| nested-panels-hit | CRAFT-03@1 | applicable | - | hit | Page section floats as a shadowed card containing three more bordered cards | Card wraps section and nested Cards provide spacing only | No modal, tool, or repeated-item semantics justify frames | Flatten section, group with spacing/dividers, retain frame only for tool |
+| drawer-tool-clear | CRAFT-03@1 | applicable | - | clear | Unframed page contains one bounded detail drawer | Drawer primitive owns focus and dismiss behavior; no nested Card | Detail drawer is a genuine framed tool | - |
+| blue-everywhere-hit | CRAFT-04@1 | applicable | - | hit | Canvas, surfaces, actions, charts, and state badges use blue variants | Single blue ramp fills surface, action, success, and selected roles | No verified monochrome baseline or semantic role separation | Restore neutral surfaces and assign distinct semantic state roles |
+| semantic-color-clear | CRAFT-04@1 | applicable | - | clear | Neutral surfaces dominate; cyan marks selection, amber warning, red failure | Tokens map each semantic role independently | Brand accent remains selective and states stay distinguishable | - |
+| pill-everything-hit | CRAFT-05@1 | applicable | - | hit | Tabs, buttons, fields, filters, and section labels all use pill silhouettes | Global radius-full token applies across unrelated controls | No baseline or semantic grouping justifies uniform pill geometry | Restore control-specific shapes and reserve pills for compact tags |
+| status-pill-clear | CRAFT-05@1 | applicable | - | clear | Compact pills appear only for status and removable filter chips | Badge and Chip primitives use pill shape; buttons use restrained radius | Shape communicates status/grouping and is not page-wide | - |
+| static-hero-not-applicable | CRAFT-01@1 | not-applicable | Planning-only review of a static hero composition; no Fill output with claimable primary action is in scope | - | - | - | No Fill surface to walk | - |
+| source-map-blocked | CRAFT-03@1 | blocked | Container nesting source map was not included in review input; rendered layer alone cannot decide framing semantics | - | Rendered shows two nested framed surfaces; nesting intent unclear | Source nesting map absent from review input | No exception can be checked without source | Include the container nesting map in review input before complete craft Pass |

package/examples/craft-detectors/existing-brand-contrast.md

@@ -1,5 +1,7 @@
 # Existing-brand baseline contrast
 
+Worked fixture for registry audit rows against a verified project baseline. Registry entries are evaluated in seven-column form; generic registry outcomes never override a verified baseline, and safety or explicit declarations override baseline consistency.
+
 ## Verified baseline exception
 
 - Detector: CRAFT-05
@@ -20,3 +22,12 @@
 - Override disposition: hit
 - Reason: safety, usability, and explicit dangerous-action declarations override baseline consistency.
 - Positive fix: retain baseline icon treatment for ordinary tools, but use explicit destructive text and confirmation for account deletion.
+
+## Craft audit
+
+| ID@ver | Applicability | Predicate reason / missing proof | Result | Rendered evidence | Source evidence | Exception check | Positive fix |
+| --- | --- | --- | --- | --- | --- | --- | --- |
+| CRAFT-05@1 | applicable | - | clear | Pill silhouette is consistent, sparse, and does not obscure control state | Confirmed project radius token and shared primitives apply the geometry consistently | Verified baseline: capsule geometry is a first-party brand convention; binding status is `ready` | - |
+| CRAFT-07@1 | applicable | - | hit | Destructive action has no text, consequence, or confirmation cue | IconButton invokes deletion directly without confirmation Dialog | safety, usability, and explicit dangerous-action declarations override baseline consistency | Retain baseline icon treatment for ordinary tools, but use explicit destructive text and confirmation for account deletion |
+| CRAFT-08@1 | not-applicable | Audited brand surfaces declare no motion; static brand audit has no motion face in scope | - | - | - | No motion face to check | - |
+| CRAFT-06@1 | blocked | Dense-container type token map was not included in the brand audit input | - | Rendered heading observed at brand display scale | Type token map absent from review input | No exception can be checked without source | Include the type token map in review input before complete craft Pass |

package/examples/craft-detectors/landing-product-contrast.md

@@ -1,12 +1,14 @@
-# Landing and product detector contrast cases
+# Landing and product craft contrast cases
 
-Authored worked cases for CRAFT-06 through CRAFT-08. Each result is advisory and leaves declaration mapping, severity, and verdict to `ui-evaluator`.
+Authored worked cases for CRAFT-06 through CRAFT-08 against the registry. Each applicable result is advisory and leaves declaration mapping, severity, and verdict to `ui-evaluator`.
 
-| Case | ID | Expected | Rendered evidence | Source evidence | Exception check | Owner hint | Positive fix |
-| --- | --- | --- | --- | --- | --- | --- | --- |
-| dashboard-display-type-hit | CRAFT-06 | hit | Compact metrics panel uses 72px heading that crowds values below | Panel heading references display-hero token | Surface is an operational panel, not a literal hero | design | Use panel title role and preserve hierarchy through grouping and weight |
-| product-hero-clear | CRAFT-06 | clear | Product name leads first viewport while next section remains visible | H1 uses display token within bounded responsive hero | Literal product hero warrants display scale and text fits target viewports | design | - |
-| verbose-toolbar-hit | CRAFT-07 | hit | Toolbar repeats Undo, Redo, Save, Zoom in, and Zoom out as wide text pills | Generic Button renders familiar tool actions despite icon library availability | Actions are familiar and low-risk; accessible icon labels are possible | components | Use icon buttons with accessible names and tooltips; keep text for ambiguous commands |
-| destructive-text-clear | CRAFT-07 | clear | Delete workspace remains an explicit text action with consequence copy | Destructive Button opens confirmation Dialog | High-risk action benefits from explicit language | components | - |
-| looping-decoration-hit | CRAFT-08 | hit | CTA badge bounces continuously without state change | Infinite keyframes animate translateY and ignore reduced-motion | No game or immersive intent is declared | craft | Remove loop and use short transform/opacity feedback only for a named state change |
-| state-transition-clear | CRAFT-08 | clear | Saved indicator fades in after persistence and stays stable | 160ms opacity transition follows saved state and reduced-motion disables it | Motion explains persistence completion | craft | - |
+| Case | ID@ver | Applicability | Predicate reason / missing proof | Result | Rendered evidence | Source evidence | Exception check | Positive fix |
+| --- | --- | --- | --- | --- | --- | --- | --- | --- |
+| dashboard-display-type-hit | CRAFT-06@1 | applicable | - | hit | Compact metrics panel uses 72px heading that crowds values below | Panel heading references display-hero token | Surface is an operational panel, not a literal hero | Use panel title role and preserve hierarchy through grouping and weight |
+| product-hero-clear | CRAFT-06@1 | applicable | - | clear | Product name leads first viewport while next section remains visible | H1 uses display token within bounded responsive hero | Literal product hero warrants display scale and text fits target viewports | - |
+| verbose-toolbar-hit | CRAFT-07@1 | applicable | - | hit | Toolbar repeats Undo, Redo, Save, Zoom in, and Zoom out as wide text pills | Generic Button renders familiar tool actions despite icon library availability | Actions are familiar and low-risk; accessible icon labels are possible | Use icon buttons with accessible names and tooltips; keep text for ambiguous commands |
+| destructive-text-clear | CRAFT-07@1 | applicable | - | clear | Delete workspace remains an explicit text action with consequence copy | Destructive Button opens confirmation Dialog | High-risk action benefits from explicit language | - |
+| looping-decoration-hit | CRAFT-08@1 | applicable | - | hit | CTA badge bounces continuously without state change | Infinite keyframes animate translateY and ignore reduced-motion | No game or immersive intent is declared | Remove loop and use short transform/opacity feedback only for a named state change |
+| state-transition-clear | CRAFT-08@1 | applicable | - | clear | Saved indicator fades in after persistence and stays stable | 160ms opacity transition follows saved state and reduced-motion disables it | Motion explains persistence completion | - |
+| plain-docs-not-applicable | CRAFT-07@1 | not-applicable | Static documentation page under review declares no repeated operation surface; no toolbar or high-frequency action face is in scope | - | - | - | No repeated-operation face to check | - |
+| motion-source-blocked | CRAFT-08@1 | blocked | Motion implementation source was not included in review input for this case; transition purpose and reduced-motion handling cannot be decided | - | Rendered shows an entrance transition on the pricing section | Motion source absent from review input | No exception can be checked without source | Provide the motion source before complete craft Pass |

package/examples/craft-detectors/saas-dashboard.md

@@ -1,29 +1,30 @@
-# SaaS dashboard detector contrast
+# SaaS dashboard craft audit
 
-Worked fixture for detector protocol validation. Rendered and source descriptions are independent authored observations, not validator-generated expectations.
+Worked fixture for the seven-column registry audit format. Rendered and source descriptions are independent authored observations, not validator-generated expectations.
 
 ## Input
 
 - Rendered surface: desktop operations dashboard at 1440 x 900.
 - Source surface: React page using shared Button, Table, Card, Badge, and motion tokens.
 - Baseline: none.
+- Registry: `skills/design-playbook/references/rules.md`, full catalog enabled (no contract subset declared).
 
-## Detector ledger
+## Craft audit
 
-| ID | Status | Rendered evidence | Source evidence | Exception check | Positive fix |
-| --- | --- | --- | --- | --- | --- |
-| CRAFT-01 | hit | Header has three equally filled primary actions | Three Button instances use `variant="primary"` | No equal-choice flow in spec | Keep Create run primary; move Import and Export to secondary/menu actions |
-| CRAFT-02 | hit | Twelve equal cards hide queue comparison | Queue records map directly to Card wrappers | Records are comparable operational rows, not independent browse objects | Use a table/list with stable columns and one framed detail tool |
-| CRAFT-03 | clear | Page bands are unframed; detail drawer alone is framed | Card is not nested and Drawer owns detail | Drawer is a genuine framed tool | - |
-| CRAFT-04 | clear | Neutral surfaces carry most area; amber and red have named state roles | Semantic tokens map warning and failure independently | No monochrome exception needed | - |
-| CRAFT-05 | clear | Pills appear only for status badges | Badge uses pill radius; buttons and inputs use control-specific shape | Status grouping warrants compact pill | - |
-| CRAFT-06 | clear | Panel headings fit dense dashboard hierarchy | Display type token is absent inside panels | No hero context claimed | - |
-| CRAFT-07 | clear | Refresh and close use named icon buttons; destructive action keeps text | IconButton has accessible labels and destructive Button remains explicit | High-risk action needs text | - |
-| CRAFT-08 | blocked | Static capture cannot prove transition purpose or reduced-motion behavior | Motion source was not included in review input | No exception can be checked without source | Provide motion source and interaction trace before complete craft Pass |
+| ID@ver | Applicability | Predicate reason / missing proof | Result | Rendered evidence | Source evidence | Exception check | Positive fix |
+| --- | --- | --- | --- | --- | --- | --- | --- |
+| CRAFT-01@1 | applicable | - | hit | Header has three equally filled primary actions | Three Button instances use `variant="primary"` | No equal-choice flow in spec | Keep Create run primary; move Import and Export to secondary/menu actions |
+| CRAFT-02@1 | applicable | - | hit | Twelve equal cards hide queue comparison | Queue records map directly to Card wrappers | Records are comparable operational rows, not independent browse objects | Use a table/list with stable columns and one framed detail tool |
+| CRAFT-03@1 | applicable | - | clear | Page bands are unframed; detail drawer alone is framed | Card is not nested and Drawer owns detail | Drawer is a genuine framed tool | - |
+| CRAFT-04@1 | applicable | - | clear | Neutral surfaces carry most area; amber and red have named state roles | Semantic tokens map warning and failure independently | No monochrome exception needed | - |
+| CRAFT-05@1 | not-applicable | Surface declares one restrained standard control radius in design tokens; no pill or shape-geometry variation face is in scope | - | - | - | No geometry face to check | - |
+| CRAFT-06@1 | applicable | - | clear | Panel headings fit dense dashboard hierarchy | Display type token is absent inside panels | No hero context claimed | - |
+| CRAFT-07@1 | applicable | - | clear | Refresh and close use named icon buttons; destructive action keeps text | IconButton has accessible labels and destructive Button remains explicit | High-risk action needs text | - |
+| CRAFT-08@1 | blocked | Motion source was not included in review input; static capture cannot prove transition purpose or reduced-motion behavior | - | Static capture shows a CTA badge mid-animation | Motion source absent from review input | No exception can be checked without source | Provide motion source and interaction trace before complete craft Pass |
 
 ## Expected contract observations
 
-- All eight stable IDs appear exactly once.
-- Ledger contains `hit`, `clear`, and `blocked`.
+- All eight registry craft IDs appear exactly once with pinned versions.
+- The ledger demonstrates `applicable` rows with `hit` and `clear`, one `not-applicable` row carrying an observable reason, and one `blocked` row naming its missing proof.
 - Hit rows carry rendered evidence, source evidence, exception check, and positive fix.
 - No row assigns declaration source, severity, or verdict.

package/examples/dogfood/README.md

@@ -0,0 +1,46 @@
+# Dogfood — S6 P3 全量档（本仓库自身 UI 面）
+
+S6 dogfood fixture（vNext 切片 6，工单 #41）：对本仓库自己的 showcase UI 面——SwarSight
+模拟运行队列监控页（[`../../showcase/01-spec.md`](../../showcase/01-spec.md) 的运行面）——
+跑一次 **P3 全量档全链**的静态工件集。
+
+- 场景：把队列监控页升级为「全局模拟运行中心」第一步——修订既有 decided `l1.goal`
+  （supersedes）+ region 构成变更（顶栏计数 → 全局运行控制台），命中 E 档判据 2 → P3。
+- 全链义务：spec 契约投影（成形会话 S0-S6 + G9）、run-profile P3、E 档 DD 条目两轮
+  preview 确认（G5/G10）、R3 重入（DD supersedes）+ 一条 R4 回流、方法语义证据
+  （manifest 五键）、注册表适用谓词全目录求值（G8 run 级，13 行）、五态×页面采样矩阵
+  完整执行（G11 S6 面：P3 档矩阵块强制）、六块报告 + invalidated 失效集、verdict Pass。
+- 性质：**静态合成**（synthesized walkthrough）——工件的证据/哈希/事务记录为演示值，
+  不依赖 chromium 与任何 provider 运行时；机器面由 `validate_run.py` 全量消费
+  （见 `tests/test_vnext_s6.py` 的 dogfood 走查）。
+
+## 走查要点（LR0-LR10）
+
+| 环节 | 工件 | 门禁 |
+| --- | --- | --- |
+| LR1 定档 | `run/plan.md` run-profile 块（tier P3） | — |
+| LR2 成形 | `run/shaping/shaping-log.jsonl` + `queue.json`；投影 D-0008..D-0011 | G9 + G1 + G7 |
+| LR3 决策 | `run/decision-report.md` DD-0001（E 档，后被 R3 挑战）/ DD-0002（supersedes 修订） | G10 |
+| LR4 预览 | `run/preview/` 两轮 decision/confirm 事务（round-1 选 B；round-2 修订后再确认 B2） | G5 |
+| LR5 实现 | `run/filled-ui.md`（round-2 修订触发 Re-Fill） | — |
+| LR6 工艺 | `run/craft-guard.md`（注册表 13 条全求值） | G8 run 级 |
+| LR7 证据 | `run/evidence/manifest.jsonl`（方法语义五键；首轮失效证据 r2 重采） | G6 |
+| LR8 评审 | `run/point-back.md` 六块 + 采样矩阵 10/10 全采样 | G2/G11 |
+| LR9 回流 | R3（dd: DD-0001 → DD-0002 supersedes）+ R4（暂停按钮可访问名）→ 失效集重评 | 轮次/G4 |
+| LR10 终局 | verdict Pass；基线漂移三出口「保持」复核 | G3/G12 |
+
+## 校验命令
+
+```bash
+python packages/design-playbook/scripts/validate_run.py \
+  packages/design-playbook/examples/dogfood/run/spec.md \
+  packages/design-playbook/examples/dogfood/run/point-back.md \
+  --preview-dir  packages/design-playbook/examples/dogfood/run/preview \
+  --decision-report packages/design-playbook/examples/dogfood/run/decision-report.md \
+  --evidence-dir packages/design-playbook/examples/dogfood/run/evidence \
+  --run-root     packages/design-playbook/examples/dogfood/run \
+  --contract-project packages/design-playbook/examples/dogfood/project \
+  --contract-run packages/design-playbook/examples/dogfood/run \
+  --shaping-dir  packages/design-playbook/examples/dogfood/run/shaping \
+  --strict
+```

package/examples/dogfood/project/DESIGN.md

@@ -0,0 +1,23 @@
+# SwarSight — DESIGN.md（dogfood 项目基线，静态合成）
+
+Provenance: observed（showcase 演示面提炼；本轮 P3 升级前绑定）。
+
+## 布局段（layout conventions）
+
+- 工作台骨架：顶栏全局操作 + 主工作区；顶栏不承载任务级数据。
+- 主列表区：可比较表格（行记录），等权卡片墙仅用于画廊型内容。
+- 分区无框：页面分区靠间距与字重；有界容器仅用于对话框与抽屉。
+- 中性面主导；语义色仅作状态命名（成功/警告/失败），不作装饰。
+- 控件几何各自特定（按钮/输入/复选不共享统一胶囊半径）。
+- 紧凑控制台层级用结构与字重表达，不引入 display 级大字号。
+- 动效仅用于状态反馈（进入/退出/进行中），永不用于装饰。
+
+## 反馈段
+
+- 进行中动作：入口 busy + 禁重复；页面级通知用 toast（含可访问名称与 role）。
+- 空态：非白屏空态 + 下一步 CTA；加载态优先条目级进度。
+
+## 命名段
+
+- 运行状态枚举：queued / running / paused / failed / completed / aborted / timeout。
+- 批量操作条：选中后浮现；破坏性动作保留显式文字。

package/examples/dogfood/project/contract.json

@@ -0,0 +1,83 @@
+{
+  "schemaVersion": 1,
+  "changelog": [
+    {
+      "at": "2026-08-14T08:35:00Z",
+      "summary": "showcase main run projection (queue monitoring)"
+    },
+    {
+      "at": "2026-08-14T09:14:00Z",
+      "summary": "P3 dogfood projection: l1.goal supersedes D-0001 via D-0008; adds l6.c4-c6; stages sim.control_scope"
+    }
+  ],
+  "fields": {
+    "l1.goal": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "把队列监控页升级为全局模拟运行中心的第一步——运行状态跨页可查、全局控制可用、失败可批量治理"
+    },
+    "l1.target_user": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "运维分析师、场景设计者"
+    },
+    "l1.non_goals": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "不在此编辑场景；不承载完整决策审计（在 Workbench）"
+    },
+    "l1.scenes": {
+      "provenance": "inferred",
+      "resolution": "assumed",
+      "value": "日常监控在跑模拟 + 失败批量重试 + 资源异常定位",
+      "notes": "理由：场景清单延续 showcase 主 run；回退：仅日常监控。"
+    },
+    "l2.entry_choice": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "顶栏计数 + 主列表 + 侧区趋势（既有三段布局）"
+    },
+    "l6.c1": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "Given 运维打开队列监控 When 查看运行列表 Then 每条模拟运行呈现状态、场景名、耗时与资源占用（证据：交互记录）"
+    },
+    "l6.c2": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "Given 存在失败运行 When 选中失败条目触发批量重试 Then 重试入队且条目状态转为 queued（证据：交互记录）"
+    },
+    "l6.c3": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "Given 运行超时 When 选择重试或中止 Then 对应动作执行且状态迁移可见（证据：交互记录）"
+    },
+    "l6.c4": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "Given 模拟运行进行中 When 离开并返回队列页 Then 全局控制台运行 feed 恢复显示且状态可续读（证据：交互记录）"
+    },
+    "l6.c5": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "Given 多运行并发 When 在全局控制台触发全局暂停 Then 进行中运行进入 paused 且控制台持续反馈（证据：交互记录）"
+    },
+    "l6.c6": {
+      "provenance": "observed",
+      "resolution": "decided",
+      "value": "Given 控制台 feed 达到上限 When 新运行入队 Then 最旧完成条目让位且失败项不静默丢失（证据：交互记录）"
+    },
+    "sim.max_parallel": {
+      "provenance": "inferred",
+      "resolution": "assumed",
+      "value": 8,
+      "notes": "理由：并发上限延续平台默认；回退：4。"
+    },
+    "sim.control_scope": {
+      "provenance": "inferred",
+      "resolution": "assumed",
+      "value": "全局暂停/恢复不作用于手动单次重试",
+      "notes": "理由：控制范围未答；风险：误伤手动重试；回退：全局暂停覆盖全部。"
+    }
+  }
+}

package/examples/dogfood/project/decisions.jsonl

@@ -0,0 +1,11 @@
+{"confirmed_at": "2026-08-14T08:30:00Z", "decision": "查看所有模拟运行的实时状态、失败重试与资源占用的队列监控页", "field": "l1.goal", "id": "D-0001", "rationale": "showcase 主 run 用户确认的监控页目标"}
+{"confirmed_at": "2026-08-14T08:30:00Z", "decision": "运维分析师、场景设计者", "field": "l1.target_user", "id": "D-0002", "rationale": "showcase 主 run L1 确认"}
+{"confirmed_at": "2026-08-14T08:30:00Z", "decision": "不在此编辑场景；不承载完整决策审计（在 Workbench）", "field": "l1.non_goals", "id": "D-0003", "rationale": "showcase 主 run 非目标确认"}
+{"confirmed_at": "2026-08-14T08:31:00Z", "decision": "Given 运维打开队列监控 When 查看运行列表 Then 每条模拟运行呈现状态、场景名、耗时与资源占用（证据：交互记录）", "field": "l6.c1", "id": "D-0004", "rationale": "showcase 主 run L6 确认"}
+{"confirmed_at": "2026-08-14T08:31:00Z", "decision": "Given 存在失败运行 When 选中失败条目触发批量重试 Then 重试入队且条目状态转为 queued（证据：交互记录）", "field": "l6.c2", "id": "D-0005", "rationale": "showcase 主 run L6 确认"}
+{"confirmed_at": "2026-08-14T08:31:00Z", "decision": "Given 运行超时 When 选择重试或中止 Then 对应动作执行且状态迁移可见（证据：交互记录）", "field": "l6.c3", "id": "D-0006", "rationale": "showcase 主 run L6 确认"}
+{"confirmed_at": "2026-08-14T08:32:00Z", "decision": "顶栏计数 + 主列表 + 侧区趋势（既有三段布局）", "field": "l2.entry_choice", "id": "D-0007", "rationale": "showcase 主 run 布局取舍"}
+{"confirmed_at": "2026-08-14T09:12:00Z", "decision": "把队列监控页升级为全局模拟运行中心的第一步——运行状态跨页可查、全局控制可用、失败可批量治理", "field": "l1.goal", "id": "D-0008", "supersedes": "D-0001", "rationale": "P3 升级：运行中心第一步（CP-A 确认，supersedes D-0001）"}
+{"confirmed_at": "2026-08-14T09:12:00Z", "decision": "Given 模拟运行进行中 When 离开并返回队列页 Then 全局控制台运行 feed 恢复显示且状态可续读（证据：交互记录）", "field": "l6.c4", "id": "D-0009", "rationale": "CP-A 确认（feed 跨页续读）"}
+{"confirmed_at": "2026-08-14T09:12:00Z", "decision": "Given 多运行并发 When 在全局控制台触发全局暂停 Then 进行中运行进入 paused 且控制台持续反馈（证据：交互记录）", "field": "l6.c5", "id": "D-0010", "rationale": "CP-A 确认（全局暂停反馈）"}
+{"confirmed_at": "2026-08-14T09:12:00Z", "decision": "Given 控制台 feed 达到上限 When 新运行入队 Then 最旧完成条目让位且失败项不静默丢失（证据：交互记录）", "field": "l6.c6", "id": "D-0011", "rationale": "CP-A 确认（feed 上限治理）"}

package/examples/dogfood/run/candidates/console-region.html

@@ -0,0 +1,26 @@
+<!doctype html>
+<html lang="zh-CN">
+<head><meta charset="utf-8"><title>candidate B — global run console region (sketch)</title></head>
+<body>
+  <!-- provider-adapter sketch (fidelity: sketch); referenced by DD-0001 candidate B -->
+  <main>
+    <section aria-label="全局运行控制台">
+      <header>
+        <span>运行中 3</span><span>已暂停 1</span><span>失败 2</span>
+        <button type="button" aria-label="全局暂停所有进行中模拟">全局暂停</button>
+        <button type="button" aria-label="恢复已暂停的模拟">恢复</button>
+      </header>
+      <ul aria-label="运行 feed（跨页保持）">
+        <li>sim-0042 · running · 62%</li>
+        <li>sim-0041 · completed · 结果入口</li>
+        <li>sim-0039 · failed · 重试</li>
+      </ul>
+    </section>
+    <section aria-label="运行列表">
+      <table>
+        <caption>模拟运行列表</caption>
+      </table>
+    </section>
+  </main>
+</body>
+</html>

package/examples/dogfood/run/contract-bind.json

@@ -0,0 +1,78 @@
+{
+  "schemaVersion": 1,
+  "ok": true,
+  "open_fields": [],
+  "stale_fields": [],
+  "blockers": [],
+  "assumed_fields": [
+    "l1.scenes",
+    "sim.control_scope",
+    "sim.max_parallel"
+  ],
+  "bound_contract": {
+    "schemaVersion": 1,
+    "changelog": [
+      {
+        "at": "2026-08-14T08:35:00Z",
+        "summary": "showcase main run projection (queue monitoring)"
+      }
+    ],
+    "fields": {
+      "l1.goal": {
+        "provenance": "observed",
+        "resolution": "decided",
+        "value": "查看所有模拟运行的实时状态、失败重试与资源占用的队列监控页"
+      },
+      "l1.target_user": {
+        "provenance": "observed",
+        "resolution": "decided",
+        "value": "运维分析师、场景设计者"
+      },
+      "l1.non_goals": {
+        "provenance": "observed",
+        "resolution": "decided",
+        "value": "不在此编辑场景；不承载完整决策审计（在 Workbench）"
+      },
+      "l1.scenes": {
+        "provenance": "inferred",
+        "resolution": "assumed",
+        "value": "日常监控在跑模拟 + 失败批量重试 + 资源异常定位",
+        "notes": "理由：场景清单延续 showcase 主 run；回退：仅日常监控。"
+      },
+      "l2.entry_choice": {
+        "provenance": "observed",
+        "resolution": "decided",
+        "value": "顶栏计数 + 主列表 + 侧区趋势（既有三段布局）"
+      },
+      "l6.c1": {
+        "provenance": "observed",
+        "resolution": "decided",
+        "value": "Given 运维打开队列监控 When 查看运行列表 Then 每条模拟运行呈现状态、场景名、耗时与资源占用（证据：交互记录）"
+      },
+      "l6.c2": {
+        "provenance": "observed",
+        "resolution": "decided",
+        "value": "Given 存在失败运行 When 选中失败条目触发批量重试 Then 重试入队且条目状态转为 queued（证据：交互记录）"
+      },
+      "l6.c3": {
+        "provenance": "observed",
+        "resolution": "decided",
+        "value": "Given 运行超时 When 选择重试或中止 Then 对应动作执行且状态迁移可见（证据：交互记录）"
+      },
+      "sim.max_parallel": {
+        "provenance": "inferred",
+        "resolution": "assumed",
+        "value": 8,
+        "notes": "理由：并发上限延续平台默认；回退：4。"
+      },
+      "sim.control_scope": {
+        "provenance": "inferred",
+        "resolution": "assumed",
+        "value": "全局暂停/恢复不作用于手动单次重试",
+        "notes": "理由：控制范围未答；风险：误伤手动重试；回退：全局暂停覆盖全部。"
+      }
+    }
+  },
+  "contract_sha": "4a11c0ffee51b7d2c93a1e0f6d8e7b5c9420ad1f3e5b7c9d1a2f4e6b8c0d2a4f",
+  "decision_log_sha": "6b22d1eef62c8e3da4fb2f1a7c9f8d6e5314be2c4d6f8a0b2c4e6d8f0a2b4c6e"
+}

package/examples/dogfood/run/craft-guard.md

@@ -0,0 +1,19 @@
+# Craft audit (队列监控升级 — P3 全量档 dogfood)
+
+Registry: `skills/design-playbook/references/rules.md`, full catalog (P3 run: 适用谓词全求值). Seven-column rows; applicability predicates evaluated per entry. 横切五条目同步求值（a11y/resp/perf applicable；i18n/sec not-applicable 附理由）。
+
+| ID@ver | Applicability | Predicate reason / missing proof | Result | Rendered evidence | Source evidence | Exception check | Positive fix |
+| --- | --- | --- | --- | --- | --- | --- | --- |
+| CRAFT-01@1 | applicable | - | clear | 主列表与控制台不并置第二主行动；全局暂停为控制台治理动作而非页面主行动 | 仅批量重试在操作条使用 primary 变体 | 单主行动符合场景 | - |
+| CRAFT-02@1 | applicable | - | clear | 运行列表与 feed 均为可比较行式条目，未卡片化 | Table 承载行记录；feed 条目固定高度行式 | 可比较运维行不适用卡片例外 | - |
+| CRAFT-03@1 | applicable | - | clear | 页面分区无框；仅对话框与抽屉有界 | 控制台/列表/侧区靠间距与字重分区 | 有界对话框是真实框工具 | - |
+| CRAFT-04@1 | applicable | - | clear | 中性面主导；语义色仅状态命名（running/paused/failed） | 状态徽标使用独立语义令牌 | 无单色系例外诉求 | - |
+| CRAFT-05@1 | applicable | - | clear | 控件几何各自特定；无泛化胶囊 | 按钮与徽标使用各自控件半径 | 无品牌几何例外诉求 | - |
+| CRAFT-06@1 | applicable | - | clear | 紧凑控制台层级用结构与字重 | 控制台与主列表共用层级令牌，无 display 字号 | 无 hero 场景主张 | - |
+| CRAFT-07@1 | applicable | - | clear | 全局暂停/恢复图标按钮带可访问名（R4 修复后） | evidence/L6.2-pause-trace.json a11y 复走查 | 破坏性动作（中止）保留显式文字 | - |
+| CRAFT-08@1 | not-applicable | 本 run 表面无装饰动效（进度反馈用条目级进度文字与 busy 态，未引入动画） | - | - | - | 无动效面可查 | - |
+| A11Y-01@1 | applicable | - | clear | 控制台 feed 条目与全局控制均有可访问名称与 role；暂停失败 toast role=alert | evidence/L6.2-pause-trace.json a11y 复走查（首轮缺口见 R4 finding，已闭合） | 无已声明豁免 | - |
+| RESP-01@1 | applicable | - | clear | 声明视口 1280x800 下主路径与控制台跨页路径可用 | 响应式行为符合声明视口组 | 契约仅声明桌面视口 | - |
+| I18N-01@1 | not-applicable | 单语控制台（zh-CN），无 i18n 声明（无 i18n.* 契约字段，L1 未声明多语言用户群） | - | - | - | 单语声明成立 | - |
+| PERF-01@1 | applicable | - | clear | 长运行有持续进度感（feed 条目级进度逐 tick 更新；全局暂停 busy 即时反馈） | evidence/L6.2-pause-trace.json 反馈序列 | 反馈相称性未承诺耗时阈值（契约无阈值声明） | - |
+| SEC-01@1 | not-applicable | 声明范围无敏感操作新增（敏感模拟参数默认脱敏沿用；全局暂停非敏感操作） | - | - | - | 无敏感面可查 | - |

package/examples/dogfood/run/decision-report.md

@@ -0,0 +1,97 @@
+# Decision report (队列监控升级为「全局模拟运行中心」第一步)
+
+P3 全量档（S6 dogfood，目标面 = showcase SwarSight 队列监控页）。顶块 = Fill 消费面；DD 条目块记录 E 档探索、R3 重入修订与基线漂移复核（G10 校验）。
+
+```text
+design-baseline: DESIGN.md sha256:2c7e9f3ab5d1e04c6a8f0b2d4e6c9a1f3b5d7e9c1a3f5b7d9e1c3a5f7b9d1e3f
+scene: ops console simulation queue (run-center step 1)
+density: console-tight
+template: global run console + table + side trends
+regions:
+  console: global run summary (running/paused/failed) + run feed (in-flight and recent, persistent across views) + global pause/resume
+  table: simulation run list (status, scenario, owner, duration, resources, actions)
+  side: failure trends, queue pressure, recent failure causes
+  toast: page-level notices (global-pause failure, retry queued)
+components:
+  global pause -> Button with accessible name and role; busy state while applying
+  run feed entry -> console item with progress; done state opens result
+  batch retry -> selection-floating action bar (explicit text on destructive actions)
+  list -> Table (not card wall)
+baseline-changes: topbar counts replaced by global run console region
+risks:
+  - console becomes dumping ground -> only simulation run feed this step; feed caps at 12 entries
+  - persisted feed leaks across sessions -> entries live for the run only (sim.control_scope assumption bounds the pause scope)
+```
+
+Do not start implementation until this report exists (or the user explicitly skips).
+
+## DD-0001 — 顶栏计数区升级的呈现构成
+
+E 档：命中判据 2（region 集合/权重重组 = 构成变化）。两案比较 → preview round 1 用户确认选 B。后被 R3 挑战（见 DD-0002）。
+
+```yaml
+id: DD-0001
+tier: explore
+question: 顶栏计数区升级的呈现构成
+status: invalidated
+constraints:
+  baseline: DESIGN.md sha256:8f4d2b61c7a3e95d0f8c2a1b6e9d4c7f3a5b8d1e0c2f4a6b8d0e2c4f6a8b0d2f
+  spec: [l1.scenes, l6.c4]
+  rules: [PERF-01@1, CRAFT-01@1]
+candidates:
+  - {id: A, source: agent, created_at: 2026-08-14T09:22:00Z, fidelity: description, summary: 顶栏轻量计数增强（计数 + 最近 3 条 toast 式通知，强化既有 topbar）, deviations: none, assets: []}
+  - {id: B, source: provider-adapter, adapter: provider-a, created_at: 2026-08-14T09:24:00Z, fidelity: sketch, summary: 启用全局 run console region（汇总 + feed + 全局控制；主列表仅禁用 + 行内微标）, deviations: none, assets: [candidates/console-region.html sha256:547cbb919d2124c2536942df8eca134e8bfcbad8ebd03c8086c7d025e0902da8]}
+comparison:
+  axes:
+    - {axis: 任务适配·运行中切页 (l1.scenes), A: 不利——切页后计数无条目可续读, B: 支持——全局一处可续读}
+    - {axis: 感知反馈相称性 (PERF-01@1), A: 部分——长运行无持续进度感, B: 支持——feed 条目级进度持续}
+    - {axis: 基线·布局段 (DESIGN.md), A: 遵循——不新增 region, B: 突破——顶栏不承载任务级数据惯例被 region 取代（呈报用户裁决）}
+  tradeoffs: "A 以轻量换跨页不可续读；B 以新 region 换全局可查（需基线漂移复核）"
+selection:
+  candidate: B
+  rationale: 切页续读 + 长运行持续进度两轴支持面更大；布局段突破经用户确认接受
+  rejected:
+    - {candidate: A, reason: 运行中切页后状态无处可续读（l1.scenes 轴不利）}
+confirmation:
+  kind: user
+  via: preview-round-1 decision_id:7c3a91e4d2b6f80a5c7e9d1b3f5a8c0e
+  confirmed_at: 2026-08-14T09:32:00Z
+supersedes: null
+stale: DESIGN.md 来源哈希漂移检出（2026-08-14T09:40:00Z）——条目引用旧绑定 sha256:8f4d2b61…
+stale_review: {exit: keep, note: 新基线布局段未改 region 惯例的其余部分——复核行：console region 约定成立且其余布局段未动；重绑 sha256:2c7e9f3ab5d1e04c6a8f0b2d4e6c9a1f3b5d7e9c1a3f5b7d9e1c3a5f7b9d1e3f（stale 解除）}
+```
+
+## DD-0002 — 运行 feed 跨页保持修订（R3 重入）
+
+R3 finding（dd: DD-0001）挑战 B 案「控制台 feed 跨页保持」假设。修订 = 新条目 supersedes DD-0001（旧条目 invalidated 保留可解析）；重入命中 E 判据 4 → 仍 E 档 → preview round 2 用户再确认。
+
+```yaml
+id: DD-0002
+tier: explore
+question: 运行 feed 跨页保持修订（R3 重入）
+status: confirmed-user
+constraints:
+  baseline: DESIGN.md sha256:8f4d2b61c7a3e95d0f8c2a1b6e9d4c7f3a5b8d1e0c2f4a6b8d0e2c4f6a8b0d2f
+  spec: [l1.scenes, l6.c4, l6.c5]
+  rules: [CRAFT-01@1]
+candidates:
+  - {id: B2, source: agent, created_at: 2026-08-14T09:44:00Z, fidelity: description, summary: B 修订——全局控制台 + feed 条目持久化（返回恢复显示，含完成态入口与上限治理）, deviations: none, assets: []}
+  - {id: A2, source: agent, created_at: 2026-08-14T09:44:00Z, fidelity: description, summary: 回退顶栏计数 + 返回后锚点定位恢复, deviations: none, assets: []}
+comparison:
+  axes:
+    - {axis: 跨视图状态闭环 (l6.c4), B2: 支持——feed 条目持久化且状态可续读, A2: 部分——锚点恢复依赖行仍在视图内}
+    - {axis: 主行动唯一性 (CRAFT-01@1), B2: 遵循——控制台不争夺主行动, A2: 遵循——顶栏计数弱化处理}
+  tradeoffs: "B2 以全局持久换控制台治理成本（上限 12 条 + 让位规则）；A2 以上下文邻近换跨页风险"
+selection:
+  candidate: B2
+  rationale: 跨视图状态闭环轴支持面大且与运行中心第一步的方向一致
+  rejected:
+    - {candidate: A2, reason: 行滚动后锚点失效——同一失败模式（跨页/滚动丢状态）}
+confirmation:
+  kind: user
+  via: preview-round-2 decision_id:9d5b2f70c4e6a8d0b2f4a6c8e0d2b4f6
+  confirmed_at: 2026-08-14T09:52:00Z
+supersedes: DD-0001
+stale: DESIGN.md 来源哈希漂移检出（2026-08-14T09:40:00Z）——条目引用旧绑定 sha256:8f4d2b61…
+stale_review: {exit: keep, note: 新基线布局段未改 region 惯例的其余部分——复核行：持久 feed 约定仍成立；重绑 sha256:2c7e9f3ab5d1e04c6a8f0b2d4e6c9a1f3b5d7e9c1a3f5b7d9e1c3a5f7b9d1e3f（stale 解除）}
+```