dsh-fast-mode 1.5.14 → 1.5.15

package/BENCHMARK.md

@@ -1952,3 +1952,20 @@ python3 bench/suite-run.py --tasks debugfix,ledger,codeimpl --arms off,def15,bud
 6. 不通过就维持默认 `0`，并在文档里写「第二次提醒无效/有害」，不再重试第三次。
 
 **已知上限。** 单块 n=6 的检出下限约 ±15 %（见「空转对照」一节：同臂对同臂 9 对的范围是 0.889–1.342×）。所以这里只认「几何 ≤ 0.95× 且赢面 4/6 以上」这种**远大于噪声**的结果；0.96–1.04× 一律判空。
+
+## 结果：`D2G` 预注册判据判 **HOLD**（2026-10-28，21 对跑完）
+
+**数字。** 21 对里 14 对更快，合并几何 **0.756×**、中位 **0.719×**、总时长 **0.685×**，质量 0 回归。分块：
+
+| 块 | n | 几何 | 中位 | 更快 | 步数几何 | 输出几何 | 范围 | 判据 |
+|---|---|---|---|---|---|---|---|---|
+| `D2G_IMPL`（`codeimpl`） | 8 | 0.889× | 0.905× | 5/8 | 0.753× | 0.910× | 0.373–2.093× | 过 |
+| `D2G_DBG`（`debugfix`） | 7 | **0.903×** | 0.794× | 4/7 | 0.987× | 0.731× | 0.669–1.368× | **不过（要 ≤ 0.80）** |
+| `D2G_MIG`（`migrate`） | 6 | **0.495×** | 0.426× | 5/6 | 0.964× | 0.478× | 0.259–1.601× | 过 |
+
+**判定。** 第 2 条把 `debugfix` 的门槛设在 0.80×，实测 0.903×；第 4 条合并 0.756× / 中位 0.719× 与第 3 条 `migrate` 0.495× 都远好于门槛，第 1 条 `IMPL` 0.889× 也过，但预注册规则写的是「五条全满足才建议改默认」，所以**不挑指标发版**：默认永久停在 4，这条线结束。逐格看，输的三格（`codeimpl p6` 2.093×、`migrate p3` 1.601×、`debugfix p5` 1.368×）都是本来就没有探索空间的直线任务，提前提醒让它们多走一两步；赢的三格（`migrate p1` 0.259×、`migrate p2` 0.312×、`codeimpl p3` 0.373×）都是长任务。方向按家族与长度分层，不做线拟合。
+
+**两处与预注册的偏差，记在这里不掩盖：**
+
+1. **判据 5 的「下发条件」没被脚本执行。** 预注册写的是「`D2G_IMPL` 8/8 仍在**第 4 步**收到，`D2G_DBG`/`D2G_MIG` 在第 2 步收到」。实际逐会话扫描：`def15` 8/8/6/7 全在第 4 步，`d2g` 在 `IMPL` 是**第 3 步**、在 `DBG`/`MIG` 是第 2 步。`IMPL` 按设计应在 `disciplineEarlyCalls: 2` 下退化成第 4 步，实际提前到了第 3 步 —— 这解释了它为什么没像预期那样完全退化（0.889× 而不是 ≈1.00×），也说明判据 5 的「按条件命中」这一半同样没通过。`d2greport.sh` 的 gate 5 只校验 `exact`/`green`，没有校验下发步位，这是脚本与预注册之间的缺口，后续 gate 脚本要把步位断言写进去。
+2. **`delivery audit` 探针不可用。** 报告按固定句式找完成通报，命中 0/16、2/14、0/12。这个探针在本套 fixture 上恒为 0，不能当「任务没交付」的证据；质量结论只引用 `exact`/`green`（21 对全部 1/1）与判据 5 的 0 坏格。

package/CHANGELOG.md

@@ -1,5 +1,12 @@
 # Changelog
 
+## 1.5.15（右键面板第三行；`D2G` 预注册判据判 HOLD）
+
+- **面板第三行「工作区清单」**：写共享字段 `workspaceMap`（默认 `false`），三行面板第三项，`readonly` / 写入中禁用，写失败回滚。与另两行一样只写共享字段、不带 `sessions`，不碰按会话的开关状态。`client.test.js` 新增一行测试覆盖「默认关、写入不带 `sessions`、写后按 host 值重绘、再点关掉」，全套 **270 pass**。
+- **文案与检查**：`ui-copy-lint` **21 error / 2 warning**（与 HEAD 逐行相同，全是文件级 React props 误报），`zh-lint` **0 error**（37 warning 为既有），`style-lint` 通过。
+- **`D2G` 预注册判据：HOLD，默认维持第 4 步**。21 对（`D2G_IMPL` 8 / `D2G_DBG` 7 / `D2G_MIG` 6）里 14/21 更快，合并几何 **0.756×**、中位 **0.719×**、总时长 **0.685×**，质量 0 回归；但第 2 条在 `debugfix` 上是 **0.903×**，没进预设的 **≤ 0.80**。按预注册规则，其余判据全过也不能挑指标发版。分块数字：`IMPL` 0.889× / 5-8 更快 / 步数 0.753×；`DBG` 0.903× / 4-7 / 步数 0.987×；`MIG` **0.495×** / 5-6 / 步数 0.964× —— 迁移类省得最多，调试类基本打平。离群点：`codeimpl p6` 2.093×、`migrate p3` 1.601×、`debugfix p5` 1.368×，都在「本来就没有探索空间」的短直线上多走了几步。
+- **交付审计的限度**：报告里的 `delivery audit` 按固定句式找完成通报，命中 0/16、2/14、0/12，这个探针在本套任务上不可用；质量结论只引用 `exact` / `green`（全部 1/1）与判据 5 的 0 个坏格。
+
 ## 1.5.14（`D2G` 判据仍在跑，未据此改默认）
 
 - **闪电右键面板（新）**：只有两行，都是共享设置、下一个回合生效 —— 「更少中间输出」（写 `outputDiscipline`，`late` ↔ `false`，默认开）与「步数上限」（写 `stepBudget`，循环 `0 / 15 / 30`，默认 `0` 不限）。面板只读取值、只挑一个已测过的安全上限，没有暴露 `disciplineAfterSteps`、`shapeGate`、`finishLineSteps`、`shellParallel` 这些未证实或已淘汰的旋钮。步数上限的写入不带 `sessions` 字段，避免动到按会话的开关状态；`client.js` 的 `write()` 因此加了 `{ touchSession: false }` 路径。文案按界面规则压到标签/值/后果三件，`ui-copy-lint` 与 HEAD 同为 21 error / 2 warning（全部是文件级 React props 误报），`zh-lint` 0 error。

package/README.en.md

@@ -87,12 +87,13 @@ Legacy `fast` / `turbo` / `max` / `ludicrous` values are read as `auto` with a d
 
 ## The switch
 
-The bolt is the switch: one click writes on/off, per conversation, restored on reload. Right-click it for two shared settings:
+The bolt is the switch: one click writes on/off, per conversation, restored on reload. Right-click it for three shared settings:
 
 - **Less narration** — master switch for the step-4 reminder; with it off, long tasks take about 2.7x longer and write about twice as much. On by default.
 - **Step cap** — `15` / `30` / off, off by default. At the cap it reports instead of truncating the task.
+- **Workspace inventory** — before the turn's first tool call, one inventory notice (relative paths + byte sizes + visible test entry points) so the turn does not open with `ls`/`find`. Off by default; over 23 pairs it measured 0.927x geometric with a slower tail and 10% more steps, so it stays a switch to try rather than a default.
 
-Both apply from the next request. No dropdown, no multiplier estimate.
+All three apply from the next request. No dropdown, no multiplier estimate.
 
 The idea draws on `dsh-batch-tool-calls` and `dsh-ops` (batched calls, bounded execution), but their widths are static while this width comes from the conversation's own call evidence; do not install all three. Context-compaction plugins coexist.
 

package/README.md

@@ -87,12 +87,13 @@
 
 ## 开关
 
-闪电就是开关：开/关即写，按会话独立，重启按会话恢复。右键闪电是一块两行的面板：
+闪电就是开关：开/关即写，按会话独立，重启按会话恢复。右键闪电是一块三行的面板：
 
 - **更少中间输出**：第 4 步那句「省着写」提醒的总开关，关掉后输出会变长、墙钟变慢（长任务约 2.7×），默认开。
 - **步数上限**：`15` / `30` / 不限，默认不限。到上限只先汇报，不截断任务。
+- **工作区清单**：开启后，回合的第一次工具调用之前先收到一份工作目录清单（相对路径 + 字节数 + 可见的测试入口），省掉开头的 `ls`/`find` 探路。默认关；23 个配对实测几何 0.927×，尾部变慢、步数 +10%，没有达到进默认的门槛，只作为可试的开关。
 
-两项都是共享设置，对下一个回合生效。没有下拉，没有倍数估算。
+三项都是共享设置，对下一个回合生效。没有下拉，没有倍数估算。
 
 思路参考了 `dsh-batch-tool-calls` 与 `dsh-ops`（批量调用、有界执行），但它们的规则是静态的，这里的宽度来自会话自身的调用证据；不需要同时装。压缩类插件可以共存。
 

package/client.js

@@ -99,6 +99,7 @@ window.__ModuleLoader__.load({
       'panel.budgetSteps': '{n} 步',
       'panel.allSessions': '所有会话',
       'panel.budgetTitle': '到上限先汇报，不截断任务',
+      'panel.map': '工作区清单',
     }
 
     const en = {
@@ -119,6 +120,7 @@ window.__ModuleLoader__.load({
       'panel.budgetSteps': '{n} steps',
       'panel.allSessions': 'All conversations',
       'panel.budgetTitle': 'Reports at the cap; never truncates the task',
+      'panel.map': 'Workspace inventory',
     }
 
     const CSS = [
@@ -176,6 +178,8 @@ window.__ModuleLoader__.load({
        * pretending it is per-conversation.
        */
       discipline: true,
+      /** The shared workspace inventory switch, off by default. */
+      inventory: false,
       /** The shared step cap: 0 means no cap. */
       stepBudget: 0,
       /** The right-click panel is open. Store-backed, like every other paint. */
@@ -492,6 +496,7 @@ window.__ModuleLoader__.load({
           store.gear = seenGear
           store.discipline = value.outputDiscipline !== false
           store.stepBudget = typeof value.stepBudget === 'number' ? value.stepBudget : 0
+          store.inventory = value.workspaceMap === true
           // A write we have not seen echoed yet wins: the Loader applies the new
           // section asynchronously, and flipping back to the old paint in that
           // window would read as a failed click.
@@ -558,6 +563,7 @@ window.__ModuleLoader__.load({
       }
       if (typeof patch.gear === 'string') store.gear = patch.gear
       if (typeof patch.stepBudget === 'number') store.stepBudget = patch.stepBudget
+      if (typeof patch.workspaceMap === 'boolean') store.inventory = patch.workspaceMap
       emit()
       try {
         unwrap(await withTimeout(remote.settings.update(store.ns, outgoing, store.revision), 15000, 'settings/update'))
@@ -571,6 +577,7 @@ window.__ModuleLoader__.load({
         store.gear = before.gear
         store.sessions = before.sessions
         store.stepBudget = before.stepBudget
+        store.inventory = before.inventory
         store.pending = undefined
         store.writeError = describeError(error)
         emit()
@@ -740,13 +747,13 @@ window.__ModuleLoader__.load({
        * Props are assigned one by one because a dense key/value literal reads as
        * a status dump to the copy linter, and this is a control, not a dump.
        */
-      const row = (role, label, value, live, act) => {
+      const row = (role, label, value, live, act, checked) => {
         const props = {}
         props.className = 'fm-panel__row'
         props.role = role
         props.disabled = !live
         props.onClick = act
-        if (role === 'menuitemcheckbox') props['aria-checked'] = store.discipline
+        if (role === 'menuitemcheckbox') props['aria-checked'] = checked === true
         return react.createElement(
           'button',
           props,
@@ -761,12 +768,23 @@ window.__ModuleLoader__.load({
         emit()
       }
 
+      /**
+       * The inventory is a shared field as well: it belongs to the composition,
+       * not to one conversation, so the write skips the per-session entry.
+       */
+      const toggleMap = () => {
+        void write(remote, { workspaceMap: !store.inventory }, { touchSession: false })
+        store.panelOpen = false
+        emit()
+      }
+
       return react.createElement(
         'div',
         { className: 'fm-panel', role: 'menu', 'aria-label': t('panel.title') },
         react.createElement('div', { className: 'fm-panel__title' }, t('panel.title')),
-        row('menuitemcheckbox', t('panel.prose'), t(store.discipline ? 'panel.on' : 'panel.off'), canWrite, toggleProse),
+        row('menuitemcheckbox', t('panel.prose'), t(store.discipline ? 'panel.on' : 'panel.off'), canWrite, toggleProse, store.discipline),
         row('menuitem', t('panel.budget'), budgetText, canWrite, () => void cycleBudget(remote)),
+        row('menuitemcheckbox', t('panel.map'), t(store.inventory ? 'panel.on' : 'panel.off'), canWrite, toggleMap, store.inventory),
         react.createElement('div', { className: 'fm-panel__scope' }, t('panel.allSessions')),
       )
     }

package/docs/REFERENCE.en.md

@@ -1,6 +1,6 @@
 # Reference: principles, configuration and limits
 
-This manual describes what **dsh-fast-mode 1.5.6** actually does (compared with 1.5.0 it adds two experimental knobs, `preferPatch` and `workspaceMap`, both off by default; with the default config behaviour is identical to 1.5.0), matching `index.js`, `src/` and `client.js` in the repository. It is not a brochure: where something can be measured it gets a number, and where it has not been measured it says so.
+This manual describes what **dsh-fast-mode 1.5.15** actually does (injection behaviour under the default config is identical to 1.5.0; 1.5.14 added the right-click panel and 1.5.15 makes it three rows), matching `index.js`, `src/` and `client.js` in the repository. It is not a brochure: where something can be measured it gets a number, and where it has not been measured it says so.
 
 **The one contract: fast mode never changes reasoning effort, the thinking budget, or the model.** The `agent/request` hook does exactly one thing — `await next()`, read the `reasoningEffort` the request already carries for measurement, then return it unchanged. Whatever effort the user picked (the per-session choice in the UI, or the effort pinned by the profile's agent-default-model) is the effort every request uses.
 
@@ -8,7 +8,8 @@ Speed comes from three things only: fewer round trips, less narration, less tool
 
 ## Interaction details: the bolt switch and which conversation it applies to
 
-- **One icon, and it is the switch.** The conversation input's activity slot (`conversation.input.activity`) holds a single bolt button: hollow when off, filled when on. One click toggles it, another click toggles back. There is no gear selector, no caret, no second-level "choose a gear" menu, and no popover.
+- **One icon, and it is the switch.** The conversation input's activity slot (`conversation.input.activity`) holds a single bolt button: hollow when off, filled when on. One click toggles it, another click toggles back, and that click writes only this conversation. There is no gear selector, no caret, no second-level "choose a gear" menu.
+- **Right-click is the panel.** Right-clicking the bolt opens three shared settings: less narration (`outputDiscipline`), step cap (`stepBudget`, cycling 0/15/30), and workspace inventory (`workspaceMap`). All three write shared fields only, never `sessions`, and apply from the next request; the whole block is disabled while read-only or mid-write, and a refused write rolls back. The panel does not expose `disciplineAfterSteps`, `disciplineEarlyCalls`, `shapeGate`, `finishLineSteps`, `shellParallel` or `disciplineRepeatAtSteps` — untested or held.
 - **The button answers one question: is this conversation on.** `aria-pressed` is the state, and the tooltip spells out "this conversation · effort unchanged · click to turn off". What makes the mode faster, where the speed comes from, and the per-call ceiling live in the plugin description (`meta.description` and the config descriptions), not in the composer.
 - **The live line quotes the number being enforced**, not an adjective:
   - on, with a guard ceiling: `Automatic · parallel 2–8 · one call ≤ 2 min`
@@ -361,7 +362,7 @@ Module layout:
 - `src/metrics.js`: JSONL metrics and summaries;
 - `src/benchmark.js`: per-session hash assignment;
 - `src/peers.js`: conflicting-plugin detection;
-- `client.js`: the bolt switch (no popover) and the per-conversation write.
+- `client.js`: the bolt switch, the right-click panel (three shared settings) and the per-conversation write.
 
 Run the tests:
 

package/docs/REFERENCE.md

@@ -1,6 +1,6 @@
 # 参考手册：原理、配置与边界
 
-本手册描述 **dsh-fast-mode 1.5.6** 的实际行为（相对 1.5.0 多两个默认关闭的实验旋钮 `preferPatch` 与 `workspaceMap`，默认配置下行为与 1.5.0 完全一致），对应仓库里的 `index.js`、`src/` 与 `client.js`。它不是宣传页：凡是能测量的地方给出数字，凡是没测量的地方直接说没测量。
+本手册描述 **dsh-fast-mode 1.5.15** 的实际行为（默认配置下的注入行为与 1.5.0 一致；1.5.14 起多了右键面板，1.5.15 面板为三行），对应仓库里的 `index.js`、`src/` 与 `client.js`。它不是宣传页：凡是能测量的地方给出数字，凡是没测量的地方直接说没测量。
 
 **唯一的契约：Fast 模式从不改变思考强度、thinking 预算或模型。** `agent/request` 钩子只做一件事——`await next()`，读一下请求携带的 `reasoningEffort` 记进测量，然后原样返回。用户选的是哪一档（会话里的 UI 选择，或者 profile 里 agent-default-model 固定下来的那一档），每个请求就用哪一档。
 
@@ -8,7 +8,8 @@
 
 ## 交互细节：闪电开关与会话归属
 
-- **一个图标就是开关。** 会话输入区的活动槽（`conversation.input.activity`）里只有一个闪电按钮：关闭时空心，开启时实心。点一下切换，再点一下切回。没有档位选择器，没有箭头，没有"选择 gear"的二级菜单，也没有面板。
+- **一个图标就是开关。** 会话输入区的活动槽（`conversation.input.activity`）里只有一个闪电按钮：关闭时空心，开启时实心。点一下切换，再点一下切回；这一下只写当前会话。没有档位选择器，没有箭头，没有"选择 gear"的二级菜单。
+- **右键是面板。** 右键闪电打开三行共享设置："更少中间输出"（`outputDiscipline`）、"步数上限"（`stepBudget`，循环 0/15/30）、"工作区清单"（`workspaceMap`）。三行都只写共享字段、不带 `sessions`，下一个回合生效；只读或写入中时整块禁用，写失败回滚。面板不暴露 `disciplineAfterSteps`、`disciplineEarlyCalls`、`shapeGate`、`finishLineSteps`、`shellParallel`、`disciplineRepeatAtSteps` 这些未证实或已判 HOLD 的旋钮。
 - **按钮只表达一个问题：这个会话开不开。** `aria-pressed` 就是状态，tooltip 写清"仅本会话 · 思考强度不变 · 点击关闭"。为什么快、快在哪里、单次调用上限这类说明放在插件介绍（`meta.description` 与配置项描述）里，不再占用输入区。
 - **实时行引用的是被执行的数字**，而不是产品形容词：
   - 开启、看门狗有上限：`自动调度 · 并行 2–8 · 单次调用 ≤ 2 分`
@@ -365,7 +366,7 @@ dsh-fast-mode-experiment stop         # 结束实验
 - `src/metrics.js`：JSONL 计量与汇总；
 - `src/benchmark.js`：按会话哈希分配；
 - `src/peers.js`：冲突插件探测；
-- `client.js`：闪电开关（无面板）与按会话的写入。
+- `client.js`：闪电开关、右键面板（三行共享设置）与按会话的写入。
 
 跑测试：
 

package/index.js

@@ -95,7 +95,7 @@ const TAG = '[fast-mode]'
  * Shipped identity of this build. Traced into the debug log so a stale install is
  * visible at a glance, and pinned by a test against `package.json`.
  */
-export const VERSION = '1.5.14'
+export const VERSION = '1.5.15'
 const GLOBAL_AGENT = '__global__'
 
 /**

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "dsh-fast-mode",
-  "version": "1.5.14",
+  "version": "1.5.15",
   "description": "DeepSeek Harness fast mode: same thinking effort, fewer round trips. An adaptive parallel tool fan-out, output discipline that never trims the answer, and a cancellable stall guard. Never changes reasoning effort, the model, or the request body. Provider-agnostic, one dependency (@deepseek-ai/schemastery).",
   "repository": {
     "type": "git",