dsh-super-code 0.3.8 → 0.3.10
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.en.md +169 -0
- package/README.md +105 -175
- package/client/index.js +594 -260
- package/cordis.patch.yml +3 -7
- package/docs/Acknowledgements.md +38 -67
- package/docs/Changelog.md +176 -88
- package/docs/Compatibility.md +133 -278
- package/docs/EvaluationResults.md +176 -150
- package/docs/UsageGuide.md +218 -223
- package/docs/en/Acknowledgements.md +76 -0
- package/docs/en/Changelog.md +176 -0
- package/docs/en/Compatibility.md +181 -0
- package/docs/en/EvaluationResults.md +240 -0
- package/docs/en/UsageGuide.md +321 -0
- package/docs/images/agent-tree-preview.png +0 -0
- package/docs/images/agent-tree.png +0 -0
- package/docs/images/install-tgz-preview.png +0 -0
- package/docs/images/install-tgz.png +0 -0
- package/docs/images/preset-settings-preview.png +0 -0
- package/docs/images/preset-settings.png +0 -0
- package/docs/images/task-status-preview.png +0 -0
- package/docs/images/task-status.png +0 -0
- package/lib/connection/approvals.d.ts +21 -0
- package/lib/connection/approvals.d.ts.map +1 -0
- package/lib/connection/approvals.js +83 -0
- package/lib/connection/approvals.js.map +1 -0
- package/lib/connection/client.d.ts +8 -0
- package/lib/connection/client.d.ts.map +1 -0
- package/lib/connection/client.js +34 -0
- package/lib/connection/client.js.map +1 -0
- package/lib/connection/guidance.d.ts +10 -0
- package/lib/connection/guidance.d.ts.map +1 -0
- package/lib/connection/guidance.js +27 -0
- package/lib/connection/guidance.js.map +1 -0
- package/lib/connection/http.d.ts.map +1 -1
- package/lib/connection/http.js +83 -16
- package/lib/connection/http.js.map +1 -1
- package/lib/connection/service.d.ts +3 -1
- package/lib/connection/service.d.ts.map +1 -1
- package/lib/connection/service.js +8 -0
- package/lib/connection/service.js.map +1 -1
- package/lib/connection/store.d.ts.map +1 -1
- package/lib/connection/store.js +2 -0
- package/lib/connection/store.js.map +1 -1
- package/lib/connection/tasks.d.ts +41 -1
- package/lib/connection/tasks.d.ts.map +1 -1
- package/lib/connection/tasks.js +154 -31
- package/lib/connection/tasks.js.map +1 -1
- package/lib/connection/types.d.ts +71 -2
- package/lib/connection/types.d.ts.map +1 -1
- package/lib/connection/types.js +5 -3
- package/lib/connection/types.js.map +1 -1
- package/lib/connection-cli.d.ts +3 -0
- package/lib/connection-cli.d.ts.map +1 -0
- package/lib/connection-cli.js +40 -0
- package/lib/connection-cli.js.map +1 -0
- package/lib/core/deletion-command.d.ts.map +1 -1
- package/lib/core/deletion-command.js +2 -1
- package/lib/core/deletion-command.js.map +1 -1
- package/lib/core/windows-runtime.d.ts +16 -0
- package/lib/core/windows-runtime.d.ts.map +1 -0
- package/lib/core/windows-runtime.js +123 -0
- package/lib/core/windows-runtime.js.map +1 -0
- package/lib/dsh/budget-settings.d.ts +2 -0
- package/lib/dsh/budget-settings.d.ts.map +1 -1
- package/lib/dsh/budget-settings.js +9 -3
- package/lib/dsh/budget-settings.js.map +1 -1
- package/lib/dsh/budgets.d.ts +2 -0
- package/lib/dsh/budgets.d.ts.map +1 -1
- package/lib/dsh/budgets.js +6 -0
- package/lib/dsh/budgets.js.map +1 -1
- package/lib/dsh/connection-approval.d.ts +5 -0
- package/lib/dsh/connection-approval.d.ts.map +1 -0
- package/lib/dsh/connection-approval.js +72 -0
- package/lib/dsh/connection-approval.js.map +1 -0
- package/lib/dsh/connection-history.d.ts +6 -0
- package/lib/dsh/connection-history.d.ts.map +1 -0
- package/lib/dsh/connection-history.js +52 -0
- package/lib/dsh/connection-history.js.map +1 -0
- package/lib/dsh/connection-host.d.ts +36 -3
- package/lib/dsh/connection-host.d.ts.map +1 -1
- package/lib/dsh/connection-host.js +97 -19
- package/lib/dsh/connection-host.js.map +1 -1
- package/lib/dsh/connection.d.ts.map +1 -1
- package/lib/dsh/connection.js +4 -1
- package/lib/dsh/connection.js.map +1 -1
- package/lib/dsh/index.d.ts.map +1 -1
- package/lib/dsh/index.js +4 -0
- package/lib/dsh/index.js.map +1 -1
- package/lib/dsh/official-windows-shell.d.ts +4 -0
- package/lib/dsh/official-windows-shell.d.ts.map +1 -0
- package/lib/dsh/official-windows-shell.js +128 -0
- package/lib/dsh/official-windows-shell.js.map +1 -0
- package/lib/dsh/windows-shell.d.ts +53 -0
- package/lib/dsh/windows-shell.d.ts.map +1 -0
- package/lib/dsh/windows-shell.js +182 -0
- package/lib/dsh/windows-shell.js.map +1 -0
- package/package.json +28 -16
- package/presets/super-code/agent.cordis.yml +3 -7
- package/runtime/windows/README.md +45 -0
- package/runtime/windows/arm64-busybox.exe.gz +0 -0
- package/runtime/windows/bash.exe.gz +0 -0
- package/runtime/windows/grep.exe.gz +0 -0
- package/runtime/windows/licenses/BusyBox-LICENSE.txt +348 -0
- package/runtime/windows/licenses/GPL-3.0.txt +674 -0
- package/runtime/windows/licenses/LGPL-3.0.txt +165 -0
- package/runtime/windows/licenses/MSYS-COPYING.LIB +165 -0
- package/runtime/windows/licenses/MSYS-COPYING.RUNTIME +674 -0
- package/runtime/windows/licenses/PCRE-LICENCE.txt +93 -0
- package/runtime/windows/licenses/ncurses-COPYING.txt +29 -0
- package/runtime/windows/manifest.json +236 -0
- package/runtime/windows/msys-2.0.dll.gz +0 -0
- package/runtime/windows/msys-iconv-2.dll.gz +0 -0
- package/runtime/windows/msys-intl-8.dll.gz +0 -0
- package/runtime/windows/msys-pcre-1.dll.gz +0 -0
- package/runtime/windows/sources/busybox-w32-FRP-6354-g5f09fb446.tgz +0 -0
- package/runtime/windows/sources.json +45 -0
- package/runtime/windows/x64-busybox.exe.gz +0 -0
package/docs/Compatibility.md
CHANGED
|
@@ -1,326 +1,181 @@
|
|
|
1
|
-
#
|
|
1
|
+
# 兼容性
|
|
2
2
|
|
|
3
|
-
|
|
3
|
+
中文 · [English](en/Compatibility.md)
|
|
4
4
|
|
|
5
|
-
|
|
5
|
+
Web 已验证 Harness `0.1.5-alpha.1` ~ `0.2.1-rc.4`,其中 `rc.4` 为下述增量验收;官方桌面已验证至 `0.2.1-rc.4`(x64 包在 Rosetta 下),NEXT 已验证 `2.0.19-next`(内置 `0.2.1-rc.2`)。
|
|
6
6
|
|
|
7
|
-
|
|
7
|
+
0.3.10 的内置 Windows Bash 另在官方 Desktop `0.2.1-rc.5` 和 NEXT `2.0.21-next`(内置 `0.2.1-rc.4`)完成 Windows 完整进程专项验收,范围见下文。
|
|
8
8
|
|
|
9
|
-
|
|
9
|
+
## 已验证的宿主
|
|
10
10
|
|
|
11
|
-
|
|
11
|
+
以下 26 个目标在 0.3.6、0.3.7、0.3.8 中均完成整组验证。
|
|
12
12
|
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
**Settings and directory-consent delta (2026-10-26):** All 243 tests, type checking, build and client syntax checks pass. Single-instance sequential checks cover Web `0.1.5-alpha.1` and `0.2.1-rc.2`, official macOS arm64 Desktop `0.2.1-rc.2`, and installed NEXT `2.0.19-next` (Harness `0.2.1-rc.2`). They verify the Home overview, separate protection settings, directory checkboxes, expanded credentials, actual MCP connections and global consent with settings closed. Approval enables native session execution over the same connection; rejection grants nothing. Screenshots were saved with no page script errors. Web checks also exercise browse-directory selection and cancellation in an SSH launch scenario. The native picker entry was invoked, but OS accessibility timed out; native selection confirmation/cancellation remains unverified. The 26-target matrix below is evidence for the previous candidate, not a full rerun of this delta.
|
|
16
|
-
|
|
17
|
-
2026-10-26 后续布局调整:关联会话及客户端详情移至“连接”,首页仅展示在线/历史对话数量;预设配置改为直接展示,去掉折叠和重复标题。245 项测试、类型检查、构建和客户端语法检查通过。最低 Web `0.1.5-alpha.1`、最高已适配 Web `0.2.1-rc.2`、官方 macOS arm64 Desktop `0.2.1-rc.2` 与 NEXT `2.0.19-next` 完成单实例串行增量验收并保存截图,页面脚本错误为零:连接页断开后会话归入历史,经新连接明确继续后恢复在线且不重复显示;幂等找回仅更新关联,不重复派发。独立浏览器组件检查另覆盖空状态和打开会话。中间版本未重跑本次布局,原生目录选择确认/取消的限制仍同上。
|
|
18
|
-
|
|
19
|
-
The subsequent layout delta on 2026-10-26 moves client and linked-session details to Connection, keeps only connected/past conversation counts on Home, and displays the preset form directly without repeated headings. All 245 tests, type checking, build and client syntax checks pass. Sequential single-instance checks with screenshots pass on minimum Web `0.1.5-alpha.1`, latest adapted Web `0.2.1-rc.2`, official macOS arm64 Desktop `0.2.1-rc.2`, and NEXT `2.0.19-next`, with no page script errors. Disconnecting from Connection moves the conversation to history; an explicit follow-up over a new connection restores its connected status without duplication. Idempotent retries update the association without dispatching again. Separate browser component checks also cover empty states and opening conversations. Intermediate versions were not rerun for this layout; the native-picker confirmation/cancellation limitation above remains.
|
|
20
|
-
|
|
21
|
-
`0.3.8` 基于正式版 `0.3.7`,合入 `feat/mcp-delegation` 的默认关闭本机 MCP 服务与连接设置。下方运行预算矩阵中的 **20 个 Web 版本、5 个官方 macOS arm64 Desktop 版本及 DSH NEXT** 均通过基础连接验收:设置保存、真实 MCP SDK 客户端认证、原生 Super Code 根会话、重复请求去重、同会话补充要求与运行中取消。宿主使用隔离 profile,不修改日常配置。
|
|
22
|
-
|
|
23
|
-
Version `0.3.8`, based on released `0.3.7`, merges the opt-in local MCP service and connection settings from `feat/mcp-delegation`. The **20 Web versions, five official macOS arm64 Desktop versions and DSH NEXT** in the runtime-budget matrix below pass baseline connection checks: settings persistence, authentication with a real MCP SDK client, native Super Code root sessions, duplicate-request coalescing, follow-ups and live cancellation. Isolated profiles leave daily configuration untouched.
|
|
24
|
-
|
|
25
|
-
早期三标签候选的增量验收覆盖最低 Web `0.1.5-alpha.1`、已适配最高 Web `0.2.1-rc.2`、官方 Desktop `0.2.1-rc.2` 和 NEXT `2.0.17-next`。四端均通过三标签设置、独立连接草稿、凭证轮换时保留草稿、切页隐藏凭证、通用“连接 Prompt”、原生子 Agent 执行和整树取消。冷进程重启保留启用设置、凭证与任务 ID;未完成任务没有自动重发或误报成功,显式补充要求后可继续同一会话。官方 Desktop `0.2.0-rc.1/rc.2` 另通过上一 UI 候选的串行增量检查,其余版本保留基础候选记录,不宣称全部复验最终文案。
|
|
26
|
-
|
|
27
|
-
Earlier three-tab candidate checks cover minimum Web `0.1.5-alpha.1`, latest adapted Web `0.2.1-rc.2`, official Desktop `0.2.1-rc.2`, and NEXT `2.0.17-next`. All four pass three-tab settings, independent drafts, draft preservation during token rotation, token hiding on tab changes, the generic Connection Prompt, native child-Agent execution and whole-tree cancellation. Cold process restarts retain enabled settings, credentials and task IDs without replaying pending prompts or reporting false success; an explicit follow-up resumes the same session. Official Desktop `0.2.0-rc.1/rc.2` also passed serial delta checks on the preceding UI candidate. Other versions retain baseline evidence rather than a claim that all final wording was rerun.
|
|
28
|
-
|
|
29
|
-
上述早期增量与重启验收严格单实例串行运行,退出前一宿主后才启动下一宿主。241 项全量测试、类型检查、构建及客户端语法检查通过,包含真实 MCP HTTP、幂等冲突、任务归属、目录边界、断线、权限等待、取消、恢复和 Git 工作区观察。测试模型为本地确定性 LlmAdapter,经过真实宿主 Agent 与工具链;没有调用外部大模型,不代表模型效果评测。Web 使用 Chromium,原生仅覆盖 macOS arm64;Windows/Linux、macOS x64、Safari 专属行为与客户端根据 Prompt 自动配置连接的能力未验收。
|
|
30
|
-
|
|
31
|
-
Those earlier delta and restart runs use one DSH instance at a time, stopping the previous host before starting the next. All 241 tests, type checking, build and client syntax checks pass, including real MCP HTTP, idempotency conflicts, ownership, directory boundaries, disconnects, permission waits, cancellation, recovery and Git workspace observations. A local deterministic LlmAdapter drives actual host Agents and tools; no external model is called and this is not an effectiveness evaluation. Web uses Chromium and native coverage is macOS arm64 only. Windows/Linux, macOS x64, Safari-specific behavior and clients’ ability to configure MCP automatically from a prompt remain unvalidated.
|
|
32
|
-
|
|
33
|
-
## 0.3.7 · 删除保护与任务状态 · Deletion Protection and Task Status · 2026-10-24
|
|
34
|
-
|
|
35
|
-
正式发布候选保留稳定记忆库,未包含开发中的版本化记忆或任务 skills。229 项测试、类型检查、构建和客户端语法检查通过。下方相同的 20 个 Web、5 个官方 macOS arm64 Desktop 与 DSH NEXT,均在真实宿主组件上通过 25 项删除保护/任务投影检查,包含旧设置、官方预设隔离、逐次确认、旧/新日志、嵌套 PTC、文件归属和任务状态不干预模型请求的回归。
|
|
36
|
-
|
|
37
|
-
The release candidate retains the stable memory library and excludes experimental versioned memory and task skills. All 229 tests, type checking, build, and client syntax checks pass. The same 20 Web versions, five official macOS arm64 Desktop versions, and DSH NEXT listed below pass 25 deletion-protection/task-projection checks each against actual host components, including legacy settings, official-preset isolation, per-call approval, old/new journals, nested PTC, file attribution, and regression checks that task feedback leaves model requests unchanged.
|
|
38
|
-
|
|
39
|
-
页面增量复验覆盖最低 Web `0.1.5-alpha.1`、已适配最高 Web `0.2.1-rc.2`、官方 Desktop `0.2.1-rc.2` 和 NEXT `2.0.17-next`。每端通过 7 项任务状态检查,涵盖全部展开、实时更新、原圆圈树、减少动态效果、长文本、失败状态、373/374 行列表的滚动/键盘访问/锚点与缩短数据;仅挂载视口附近行。每端另通过 6 项删除保护设置检查,包含三档切换、独立官方开关、保存重载、旧布尔配置和恢复默认。
|
|
40
|
-
|
|
41
|
-
Incremental UI checks cover minimum Web `0.1.5-alpha.1`, latest adapted Web `0.2.1-rc.2`, official Desktop `0.2.1-rc.2`, and NEXT `2.0.17-next`. Each passes seven task-status checks covering expansion, live updates, the circular graph, reduced motion, wrapped text, failures, and scrolling/keyboard access/anchors/shrinking data across 373/374 rows while mounting only nearby rows. Each also passes six deletion-settings checks covering modes, independent official opt-in, save/reload, legacy booleans, and reset.
|
|
42
|
-
|
|
43
|
-
任务状态页面使用真实宿主 API 和构造日志,不新增模型调用,不将夹具文件记录当作真实文件编辑。运行环和暗色主题 tokens 已检查,不宣称所有宿主主题设置组合或十万行真实浏览器性能已测。严格删除确认的引擎检查由测试驱动器答复真实宿主审批服务,不等同人工点击原生确认对话框。NEXT 测试外壳已知的两条 ready 诊断另列,任务页面插件异常和其他控制台错误/警告为零。未重跑未变的安装器或扩展平台范围;中间版本页面沿用既有验证记录,Windows/Linux 原生客户端仍未验收。
|
|
44
|
-
|
|
45
|
-
Task-status UI uses real host APIs with constructed journals, with no new model calls or claims of real file edits. Running indicators and dark-theme tokens are checked; this is not every host theme-setting combination or a 100,000-row browser benchmark. Strict-deletion engine checks answer the real host approval service through the driver, not manual clicks in native approval dialogs. Two known NEXT test-shell ready diagnostics are recorded separately; task pages have no plugin exceptions or other console errors/warnings. Unchanged installers and additional platforms were not retested; intermediate-version pages retain prior evidence, and native Windows/Linux clients remain unvalidated.
|
|
46
|
-
|
|
47
|
-
## 0.3.6 · 旧预算默认配置 · Legacy Budget Defaults · 2026-10-20
|
|
48
|
-
|
|
49
|
-
0.3.6 默认使用此前 Pro 测试的旧预算:完整上下文窗口、80% 压缩触发、16% 原文保留、8,192 Token 摘要、压缩追加与溢出重试各 1 次;提供新版 PTC 的宿主默认/上限为 600 秒,仍受宿主上限约束。用户保存值和加载配置优先,运行中的 Agent 保留快照,恢复默认会回到旧预算。括号内宿主默认值提示保持不变,官方预设仍默认不应用。
|
|
50
|
-
|
|
51
|
-
Version 0.3.6 uses the older Pro-test budgets by default: full-context-window accounting, compression at 80%, 16% verbatim retention, an 8,192-token summary cap, and one additional compression attempt and overflow retry. Hosts with the newer PTC provider use 600-second default/maximum deadlines within the host ceiling. Saved values and plugin configuration take precedence; running Agents retain their snapshots, and reset restores the older budgets. Host-default hints in parentheses remain unchanged, and official presets remain opt-in.
|
|
52
|
-
|
|
53
|
-
对下方同一组 **20 个 Web、5 个官方 macOS arm64 Desktop 和 DSH NEXT** 完成默认策略的增量复验,每个目标通过 6 项页面/实际 Agent 检查和 1 项进程冷重启检查,135 个运行文件与新候选逐字节一致。195 项测试、类型、构建和客户端语法检查通过;20 个实际 Web 宿主与 NEXT 各 11 项引擎检查通过。旧宿主无 `ptcRuntime` 时不会补入不支持的时限字段。客户端源文件字节未变。
|
|
54
|
-
|
|
55
|
-
The same **20 Web versions, five official macOS arm64 Desktop versions, and DSH NEXT** listed below passed incremental checks for the new default policy: six UI/actual-Agent checks and one cold-process-restart check per target, with all 135 runtime files matching the new candidate. All 195 tests, type checking, build, and client syntax checks passed, along with 11 engine checks on each actual Web host and NEXT. Older hosts without `ptcRuntime` receive no unsupported deadline fields. The client source remains byte-identical.
|
|
56
|
-
|
|
57
|
-
本次复制此前隔离 profile 并加载新候选,未重跑未变的安装器或全部语言/冲突/布局组合;下方完整矩阵保留为上一候选的历史证据。平台限制不变,测试使用本地确定性模型,无外部模型调用或新增 SWE-bench 成绩。上述验收候选标记为 `0.3.5`;发布包改为 `0.3.6`,运行文件仅更新包版本与两种预设版本标记,其余运行内容逐文件核对一致。
|
|
58
|
-
|
|
59
|
-
This delta run cloned prior isolated profiles and loaded the new candidate. It did not rerun the unchanged installer or every language/conflict/layout combination; the complete matrix below remains evidence for the preceding candidate. Platform limits are unchanged. Tests use a local deterministic model with no external model calls or new SWE-bench scores. The tested candidate was labeled `0.3.5`. The `0.3.6` release changes only the package version and both preset-version markers; all other runtime content is verified file by file as identical.
|
|
60
|
-
|
|
61
|
-
## 运行预算 · Runtime Budgets · 2026-10-19
|
|
62
|
-
|
|
63
|
-
页面优化后已重新打包,对此前支持的 **20 个官方 Web 版本、5 个官方 macOS arm64 Desktop 版本及 DSH NEXT** 完成同一候选包复验。每个目标均通过 11 项页面功能检查、6 项真实会话检查和 3 项进程冷启动检查;安装副本的 133 个运行文件与冻结候选逐字节一致。该轮测试使用当时尚未发布、版本标记为 `0.3.5` 的开发候选。
|
|
64
|
-
|
|
65
|
-
After the settings-page changes, the same packaged candidate passed revalidation on **20 supported official Web versions, five official macOS arm64 Desktop versions, and DSH NEXT**. Each target passed 11 UI behavior checks, six real-conversation checks, and three cold-process-restart checks. All 133 installed runtime files match the frozen candidate. That run used the then-unreleased development candidate labeled `0.3.5`.
|
|
66
|
-
|
|
67
|
-
| 宿主 / Host | 本轮通过版本 / Versions revalidated |
|
|
68
|
-
|---|---|
|
|
69
|
-
| 官方 Web / Official Web | `0.1.5-alpha.1`、`alpha.2`、`rc.1`、`rc.2`、`rc.3`;`0.1.6-alpha.1`、`alpha.2`;`0.1.7-alpha.1`、`alpha.2`、`rc.1`、`rc.2`;`0.2.0-rc.1`、`rc.2`;`0.2.1-alpha.1`、`alpha.2`、`alpha.3`、`alpha.4`、`alpha.5`、`rc.1`、`rc.2` |
|
|
70
|
-
| 官方 Desktop / Official Desktop | `0.1.7-rc.1`、`0.1.7-rc.2`、`0.2.0-rc.1`、`0.2.0-rc.2`、`0.2.1-rc.2` |
|
|
71
|
-
| DSH NEXT | `2.0.17-next`(内置 Harness / bundled Harness `0.2.0-rc.2`)|
|
|
72
|
-
|
|
73
|
-
全部 20 个 Web 在 800px 宽窗口追加控件边界、滚动可见性和指针命中检查,均通过。440px 宽时宿主固定侧栏仅给插件留下 156px 内容区,不计入可用性通过范围;最低、最高版不加载插件的通用设置也存在相同限制。原先仅检查内部 scrollWidth 的第 12 项已剔除,避免误称窄窗口可用。
|
|
74
|
-
|
|
75
|
-
All 20 Web versions also pass control-boundary, scroll-visibility, and pointer-hit checks at an 800px viewport. At 440px the host navigation leaves only 156px for plugin content; narrow-window usability is not accepted. Plugin-free baselines on the earliest and latest hosts show the same settings-layout limitation. The original internal-scrollWidth check is excluded from the acceptance count.
|
|
76
|
-
|
|
77
|
-
覆盖“预设状态 / 上下文配置”标签页、按需读取、键盘切换、草稿保留、实际宿主默认值和单位、自定义选择器、保存、无效输入、并发修改冲突、重新读取、恢复默认、中英文、重载与冷启动持久化。官方预设开关默认关闭,开启与关闭后的新旧 Agent 预算隔离通过真实 Standard 会话验证,Super Code 的任务容量限制通过实际 PTC 工具调用验证;标准、PTC、极简和创造四种预设的选择、压缩、并发及卸载边界另由每版 10 项引擎专项覆盖。
|
|
78
|
-
|
|
79
|
-
Checks cover the Preset status / Context settings tabs, lazy reads, keyboard navigation, retained drafts, actual host defaults and units, custom selectors, save, invalid input, concurrent-edit conflicts, reload, reset, Chinese/English, and cold-start persistence. The official-preset switch defaults off. Real Standard conversations verify opt-in and creation-time snapshots; actual PTC calls verify Super Code task-capacity limits. Ten engine checks per Web version separately cover selection, compaction, concurrency, and disposal for Standard, PTC, Minimal, and Create.
|
|
80
|
-
|
|
81
|
-
本轮修复语言偏好尚未保存时插件提前重载、导致语言回退的问题。192 项测试、类型检查、构建及客户端语法检查通过;20 个 Web 宿主各 10 项压缩专项、NEXT 10 项压缩专项通过。15 个 Web 宿主及 NEXT 的实际 Node PTC 均验证默认预算与显式较短时限;`0.1.5` 五版没有新版 `ptcRuntime`,页面正确禁用这些时限字段,其余预算和旧版 run_code 正常使用。
|
|
82
|
-
|
|
83
|
-
A fix waits for the accepted language preference before preset synchronization can reload the page, preventing optimistic language changes from being lost. All 192 tests, type checking, build, and client syntax checks pass, along with ten compaction checks on each Web version and ten on NEXT. Actual Node PTC tests on 15 Web hosts and NEXT verify the plugin default and an explicit shorter deadline. The five `0.1.5` hosts lack the newer `ptcRuntime`; their deadline fields are correctly disabled while other budgets and legacy run_code remain usable.
|
|
84
|
-
|
|
85
|
-
官方桌面使用应用内插件页安装、启用,未修改日常应用或配置。测试使用本地确定性模型,没有外部模型调用或新增 SWE-bench 成绩。原生实测范围为 macOS arm64,Web 使用 Chromium;未覆盖原生 Windows/macOS x64 或 Safari 专属行为。逐宿主真实官方会话为 Standard,专项组合测试不代表所有内置工具、权限和终端场景均已逐一实测。
|
|
86
|
-
|
|
87
|
-
Official Desktop packages were installed and enabled through their native plugin pages without changing daily apps or profiles. A local deterministic model supplied responses, with no external model calls or new SWE-bench scores. Native testing covers macOS arm64 and Web testing uses Chromium; Windows/macOS x64 native apps and Safari-specific behavior were not checked. Per-host real official-preset conversations use Standard; the engine composition checks do not establish every built-in tool, permission, or terminal scenario.
|
|
88
|
-
|
|
89
|
-
## 0.3.5 · Harness 0.2.1-rc.2 · 2026-10-18
|
|
90
|
-
|
|
91
|
-
兼容范围扩展至 `0.2.1-rc.2`,最低保持 `0.1.5-alpha.1`。本次基于主干 0.3.4,逐版固定并审计宿主依赖,在独立 profile 中真实安装和运行。现有 peer 依赖范围已覆盖这些预发布版本;补充 v4 日志迁移检查后,修复旧插件来源记录的识别及投影缓存失效,未增加依赖,发布版本为 0.3.5。
|
|
92
|
-
|
|
93
|
-
Compatibility extends through `0.2.1-rc.2`, with the minimum remaining `0.1.5-alpha.1`. Based on main at plugin version 0.3.4, candidates were installed and run in isolated profiles, with each host's dependency graph pinned and audited. Existing peer ranges already cover these prereleases. Additional v4 migration checks identified and fixed legacy plugin-source recognition and projection-cache invalidation without adding dependencies. The release version is 0.3.5.
|
|
94
|
-
|
|
95
|
-
| Harness | Web/CLI 与真实浏览器 / Web/CLI and real browser | 原生 Skill 加载 / Native Skill loading |
|
|
13
|
+
| 宿主 | 版本 | 预设机制 |
|
|
96
14
|
|---|---|---|
|
|
97
|
-
| `0.
|
|
98
|
-
| `0.
|
|
99
|
-
| `0.
|
|
100
|
-
| `0.2.
|
|
101
|
-
| `0.2.1-rc.2` |
|
|
102
|
-
|
|
103
|
-
|
|
104
|
-
|
|
105
|
-
|
|
106
|
-
|
|
107
|
-
|
|
108
|
-
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
|
|
112
|
-
|
|
113
|
-
|
|
114
|
-
|
|
115
|
-
截至本次检查,官方 npm `latest` 为 `0.2.0-rc.2`、`alpha` 为 `0.2.1-alpha.5`、`next` 为 `0.2.1-rc.2`;桌面 nightly 同为 `0.2.1-rc.2`。[官方发布说明](https://github.com/deepseek-ai/deepseek-harness/releases/tag/dsh-v0.2.1-rc.2)。本次模型响应由本地确定性程序提供,没有调用外部模型、没有新增 SWE-bench 成绩,也不代表 Windows、macOS x64 或其他新增版本的原生桌面已复验;其他版本保留以下历史记录。
|
|
116
|
-
|
|
117
|
-
At this check, official npm tags point to `latest: 0.2.0-rc.2`, `alpha: 0.2.1-alpha.5`, and `next: 0.2.1-rc.2`; Desktop nightly is also `0.2.1-rc.2`. [Official release notes](https://github.com/deepseek-ai/deepseek-harness/releases/tag/dsh-v0.2.1-rc.2). A local deterministic fixture supplied model responses, without external model calls or new SWE-bench scores. This does not establish native Desktop revalidation on Windows, macOS x64, or the other newly covered versions; earlier records remain below.
|
|
118
|
-
|
|
119
|
-
### v4 会话日志补充验证 · 2026-10-18
|
|
120
|
-
|
|
121
|
-
物理日志解压、格式迁移、分页与游标由 Harness 的 Session 和官方 journal API 处理;插件不直接解析用户的日志文件。v4 会把旧的 `{kind: "plugin", plugin: "…"}` 来源迁移为 `{kind: "plugin:…"}`。原读取器会跳过其中的任务记录,导致升级后的任务状态、历史与归档不可见。现已识别新旧包名对应的精确来源,同时保留原有直接来源和旧包装形式;任务投影版本由 1 提升到 2,使此前遗漏记录的缓存重新回放。无需为此依赖实验性的 plugin-record API,仍兼容旧宿主。
|
|
122
|
-
|
|
123
|
-
Harness Session and official journal APIs own physical decompression, migration, pagination, and cursors; the plugin does not parse users' log files directly. V4 migrates `{kind: "plugin", plugin: "…"}` to `{kind: "plugin:…"}`. The previous reader skipped these task records, hiding task state, history, and archives after an upgrade. It now recognizes the exact current and legacy package sources while preserving direct sources and old wrappers. The task projection version increases from 1 to 2 to replay caches that previously missed those records. This does not require the experimental plugin-record API or remove older-host support.
|
|
124
|
-
|
|
125
|
-
使用官方 rc.2 格式迁移器与真实 v3 会话(含历史子会话事实)验证直接来源、新包名旧包装、旧包名旧包装三条路径;对照原生 v4 日志,核对任务状态、11 对工具证据、9 条记忆使用记录、按模型用量和历史详情。另验证旧空缓存失效、全量重建及后续尾部回放,并增加任务保存、归档、聚焦与无关插件隔离回归;173 项测试、类型检查、构建和客户端语法检查通过。本补充不代表已遍历所有历史损坏或中断日志;上方五版和原生桌面界面记录来自首轮候选包。
|
|
126
|
-
|
|
127
|
-
The official rc.2 format migrator restored real v3 sessions with historical child facts for direct sources and wrappers using both current and legacy package names. Comparisons with native v4 logs cover task state, 11 tool-evidence pairs, nine memory-usage rows, per-model usage, and historical details. Checks also cover invalidating stale empty caches, full replay, and subsequent tail replay. Added regressions cover task saving, archiving, focus, and unrelated-plugin isolation. All 173 tests, type checking, build, and client syntax checks passed. This supplement does not cover every damaged or interrupted historical log; the five-version and native Desktop UI records above used the initial candidate.
|
|
128
|
-
|
|
129
|
-
修复后的安装包另在 `0.1.5-alpha.1` 和 `0.2.1-rc.2` 中通过真实安装、激活、主/子 Agent、PTC、记忆归属、工作区隔离和冷重启历史读取回归;实际安装的任务投影实现与候选包一致。本补充未重跑浏览器及原生桌面界面,客户端仅更正日志格式注释。
|
|
130
|
-
|
|
131
|
-
The fixed package also passed real installation, activation, main/child Agents, PTC, memory attribution, workspace isolation, and historical reads after a cold restart on `0.1.5-alpha.1` and `0.2.1-rc.2`. The installed task projection matches the candidate. This supplement did not rerun browser or native Desktop UI checks; the client change only corrects a log-format comment.
|
|
132
|
-
|
|
133
|
-
## 0.3.4 · 2026-10-15
|
|
134
|
-
|
|
135
|
-
本次更新评测文档、公开结果快照与致谢,宿主依赖和运行行为不变。构建后 112 个运行文件与 0.3.3 发布包逐字节一致,两份预设组合仅将版本标记更新为 0.3.4;此前 0.3.3 的宿主验收记录继续适用,本轮未重跑 Web / 原生 Desktop 界面,也未新增模型效果评测。
|
|
136
|
-
|
|
137
|
-
This release updates evaluation documentation, the public results snapshot, and acknowledgements, with unchanged host dependencies and runtime behavior. After building, 112 runtime files are byte-for-byte identical to the 0.3.3 release package; the two preset compositions only update their version marker to 0.3.4. The 0.3.3 host-validation records below remain applicable. Web/native Desktop UI checks and new model-quality evaluations were not rerun for this release.
|
|
138
|
-
|
|
139
|
-
171 项测试、类型检查、构建及客户端语法检查通过;另核对评测快照汇总、文档链接、版本一致性和打包范围。
|
|
140
|
-
|
|
141
|
-
All 171 tests, type checking, build, and client syntax checks passed. Additional checks cover evaluation-snapshot aggregates, documentation links, version consistency, and package contents.
|
|
142
|
-
|
|
143
|
-
## 0.3.3 · 2026-10-14
|
|
144
|
-
|
|
145
|
-
支持下限保持 `0.1.5-alpha.1`,依赖范围不变。本次核对官方 npm `latest`/`next` 为 `0.2.0-rc.2`,`alpha` 为 `0.2.1-alpha.2`。以下三个宿主使用固定并审计的依赖和隔离 profile,真实安装 0.3.3 组合候选包;其他版本保留此前逐版记录,本次没有全部重测。
|
|
146
|
-
|
|
147
|
-
The minimum remains `0.1.5-alpha.1`, with unchanged dependency ranges. Official npm `latest`/`next` is `0.2.0-rc.2` and `alpha` is `0.2.1-alpha.2` on this date. The three pinned, audited hosts below installed the combined 0.3.3 candidate in isolated profiles. Other versions retain earlier per-version records and were not all rerun.
|
|
148
|
-
|
|
149
|
-
| Harness | 预设机制 / Presets | 0.3.3 Web/CLI 与浏览器 / Web/CLI and browser |
|
|
150
|
-
|---|---|---|
|
|
151
|
-
| `0.1.5-alpha.1` | 目录 / Directory | 通过 / Pass |
|
|
152
|
-
| `0.2.0-rc.2` | 声明式 / Declarative | 通过 / Pass |
|
|
153
|
-
| `0.2.1-alpha.2` | 声明式 / Declarative | 通过 / Pass |
|
|
154
|
-
|
|
155
|
-
三版均通过安装/激活、主/子 Agent、项目/全局记忆、跨工作区隔离和重启读取。实际浏览器验证覆盖单个 Super Code 入口、默认执行区、互斥展开与键盘/焦点、整行标题、同页及重复记忆定位、树分支与历史筛选恢复、记忆新增/编辑/取消/保存/删除确认、草稿及进行中保存跨区恢复;另验证运行与完成动画、减弱动态效果、320/400/560px 与极限缩放时节点/文字/连线间距。稳定版与 alpha 版另实际切换中英文、明暗主题,无页面或控制台错误。模型由本地确定性验收程序提供,没有调用外部模型。
|
|
156
|
-
|
|
157
|
-
All three passed installation/activation, main/child Agents, project/global memory, workspace isolation, and restart reads. Real-browser checks covered the unified Super Code entry, the default tree, mutually exclusive sections, keyboard/focus behavior, full-width headings, repeated same-page memory navigation, retained tree/history filters, memory creation/editing/cancellation/saving/confirmed deletion, and drafts/in-flight saves across section switches. Checks also covered running/completed animation, reduced motion, node/label/connector spacing at 320/400/560px and zoom limits. Stable and alpha hosts additionally passed Chinese/English and light/dark switching, without page or console errors. A local deterministic fixture supplied model responses; no external model was called.
|
|
158
|
-
|
|
159
|
-
内嵌前端 Skill 从同一候选包验证三版宿主的六条原生加载路径,覆盖目录预设及复制安装、声明式组合及持久化 profile、摘要目录、按需正文/参考读取和项目/用户同名覆盖。安装与加载验证不代表模型前端作品质量已经提升。
|
|
160
|
-
|
|
161
|
-
The bundled frontend Skill passed six native loading paths from the same candidate across these hosts: directory presets/copied installs, declarative bundles/persisted profiles, summary catalogs, on-demand content/reference reads, and project/user overrides. Installation and loading checks do not establish improved model-generated frontend quality.
|
|
162
|
-
|
|
163
|
-
官方 macOS arm64 Desktop `0.2.0-rc.2` 在隔离应用及 profile 中通过原生插件页面从 0.3.2 升级到 0.3.3,重启后旧预设仍为 0.3.2,差异提示与确认取消通过;确认重装及宿主自动重载后,预设版本更新为 0.3.3 并包含内嵌 Skill。统一侧栏切换、草稿恢复与完整记忆维护流程通过并留有截图;真实主会话的运行/完成动画、切换恢复、宽度/缩放、减弱动态效果及节点间距通过,不将主会话检查视为原生 live 子 Agent 工具验收。副本保留原始 CFBundleName,隔离应用身份、profile 和 userData;正式应用及日常配置未修改。
|
|
164
|
-
|
|
165
|
-
Official macOS arm64 Desktop `0.2.0-rc.2` upgraded from 0.3.2 to 0.3.3 through its native plugin page in an isolated app/profile. After restart, the old preset retained 0.3.2, with a mismatch notice and cancellable confirmation. Confirmed reinstallation and the host's automatic reload updated the preset to 0.3.3 with the bundled Skill. Unified-section switching, draft recovery, and the complete memory maintenance flow passed with screenshots. A live main session passed running/completed animation, section recovery, width/zoom, reduced motion, and node spacing; this does not constitute native live child-Agent tool validation. The copy retains CFBundleName and isolates application identity, profile, and userData; the official app and daily configuration were not modified.
|
|
166
|
-
|
|
167
|
-
171 项全量测试、类型检查、构建、客户端语法及 diff 检查通过。其他原生桌面版本和平台沿用历史覆盖,不宣称本轮全量复验;没有新增 SWE-bench 效果成绩。
|
|
168
|
-
|
|
169
|
-
All 171 tests, type checking, build, client syntax, and diff checks passed. Other native desktop versions and platforms retain historical coverage rather than a claim of full revalidation. This release adds no new SWE-bench performance scores.
|
|
170
|
-
|
|
171
|
-
## 0.3.2 · 2026-10-11
|
|
172
|
-
|
|
173
|
-
支持下限保持 `0.1.5-alpha.1`,没有新增依赖或宿主私有 UI 接口。官方 npm `latest`/`next` 仍为 `0.2.0-rc.2`,`alpha` 为 `0.2.1-alpha.2`。本次使用固定并审计的依赖和隔离 profile,在以下三个宿主重新安装 0.3.2 候选包并验收;中间版本保留下面 0.3.1 的逐版记录,本次没有全部重测。
|
|
174
|
-
|
|
175
|
-
The minimum remains `0.1.5-alpha.1`, without new dependencies or private host UI APIs. Official npm `latest`/`next` remains `0.2.0-rc.2`, and `alpha` is `0.2.1-alpha.2`. This release installed the 0.3.2 candidate in the three pinned, audited hosts below using isolated profiles. Intermediate versions retain the per-version 0.3.1 records below and were not all rerun.
|
|
176
|
-
|
|
177
|
-
| Harness | 预设机制 / Presets | 0.3.2 Web/CLI 与浏览器 / Web/CLI and browser |
|
|
178
|
-
|---|---|---|
|
|
179
|
-
| `0.1.5-alpha.1` | 目录 / Directory | 通过 / Pass |
|
|
180
|
-
| `0.2.0-rc.2` | 声明式 / Declarative | 通过 / Pass |
|
|
181
|
-
| `0.2.1-alpha.2` | 声明式 / Declarative | 通过 / Pass |
|
|
182
|
-
|
|
183
|
-
三版均通过插件安装/激活、主/子 Agent、任务与项目/全局记忆、跨工作区隔离和重启读取。浏览器检查覆盖圆形执行树、缩放拖动、320/400/560px 单行操作、上方名称气泡、自定义时间/排序菜单的鼠标与键盘操作、紧凑 Agent 详情、旧/新版真实工具卡、主题目录/关联条目/搜索、展开设置、版本、刷新及确认列表移除/取消。没有浏览器控制台或页面错误。稳定版另实际切换中英文和明暗主题,核对默认名称、版本和气泡并截图。模型由本地确定性验收程序提供,没有调用外部模型。
|
|
184
|
-
|
|
185
|
-
All three passed installation/activation, main/child Agents, task/project/global memory, workspace isolation and restart reads. Browser checks covered circular trees, zoom/pan, one-row actions at 320/400/560px, above-button name bubbles, custom time/sort menus with mouse and keyboard, compact Agent details, real tool cards from old/new journals, topic navigation/related entries/search, expanded settings, versions, refresh and removable/cancellable confirmation lists. No console or page errors occurred. The stable host additionally passed actual Chinese/English and light/dark switching with default-name, version and tooltip screenshots. A local deterministic fixture supplied model responses; no external model was called.
|
|
186
|
-
|
|
187
|
-
官方 macOS arm64 Desktop `0.2.0-rc.2` 的隔离应用副本实际验收了圆形树、命名操作、自定义菜单、真实工具详情、主题记忆库、版本和确认重装;移除后确认禁用、Escape 保留设置页、实际确认成功及恢复备份均已核对。副本保留原始 CFBundleName,隔离 bundle ID、profile 和 userData。后续启动语言等待修复由三版 Web 和单元测试复验;原生控制接口恢复时超时,未补录该最终修复的原生窗口截图。其他桌面版本和平台沿用历史范围,不宣称本次已全部验收。
|
|
188
|
-
|
|
189
|
-
An isolated copy of official macOS arm64 Desktop `0.2.0-rc.2` passed actual UI checks for circular trees, named actions, custom menus, real tool details, topic memory, versions and confirmed reinstallation. Removal disabled confirmation, Escape retained Settings, and actual reinstallation and recovery backups were checked. The copy preserves CFBundleName and isolates its bundle ID, profile and userData. The later startup-language wait fix passed all three Web hosts and unit tests; desktop-control recovery timed out, so a final native-window screenshot for that fix was not obtained. Other desktop versions and platforms retain historical coverage rather than a claim of full revalidation.
|
|
190
|
-
|
|
191
|
-
158 项测试、类型检查、构建与客户端语法检查通过。预设安装、归属校验、用户字段保护和备份逻辑未改;启动同步通过宿主公开语言偏好服务等待加载,并按宿主版本声明 `settingsScope` 或 `configForms` 依赖。
|
|
192
|
-
|
|
193
|
-
All 158 tests, type checking, build and client syntax checks passed. Preset installation, ownership checks, user-field protection and backup logic are unchanged. Startup synchronization waits through the host's public language-preference service and declares the host-appropriate `settingsScope` or `configForms` dependency.
|
|
194
|
-
|
|
195
|
-
## 0.3.1 · 2026-10-10
|
|
196
|
-
|
|
197
|
-
支持下限保持 `0.1.5-alpha.1`。截至本次核对,官方 npm 的 `latest`/`next` 指向 `0.2.0-rc.2`,`alpha` 指向 `0.2.1-alpha.2`;官方 macOS arm64 Desktop 更新源指向 `0.2.0-rc.2`。
|
|
198
|
-
|
|
199
|
-
The minimum remains `0.1.5-alpha.1`. On this date, official npm `latest`/`next` points to `0.2.0-rc.2` and `alpha` to `0.2.1-alpha.2`; the official macOS arm64 Desktop update feed points to `0.2.0-rc.2`.
|
|
200
|
-
|
|
201
|
-
下表 15 个官方 Harness 版本均真实安装 0.3.1 候选包,使用固定并审计的宿主依赖和隔离 profile。预设检查覆盖缺失安装、确认流程所用的改名/改标识符重装接口、冲突保护、语言同步和重启后读取/重装。主/子 Agent、任务与项目/全局记忆、跨工作区隔离、重启读取及网页执行树、缩放拖动、侧栏详情、记忆库、展开设置、版本展示和确认列表移除/取消均通过。模型由本地确定性测试提供,没有调用外部模型。
|
|
202
|
-
|
|
203
|
-
All 15 official Harness versions below installed the 0.3.1 candidate with pinned, audited host dependencies and isolated profiles. Preset checks covered missing installation, rename/identifier reinstallation, conflicts, language synchronization and restart recovery. Main/child Agents, task/project/global memory, workspace isolation and restart reads passed, along with web tree/zoom/pan, sidebar details, memory library, expanded settings, independent versions, and removable/cancellable confirmation lists. Tests used a local deterministic model without external model calls.
|
|
204
|
-
|
|
205
|
-
| Harness | 预设机制 / Presets | 0.3.1 Web/CLI 与浏览器 / Web/CLI and browser |
|
|
206
|
-
|---|---|---|
|
|
207
|
-
| `0.1.5-alpha.1` | 目录 / Directory | 通过 / Pass |
|
|
208
|
-
| `0.1.5-alpha.2` | 目录 / Directory | 通过 / Pass |
|
|
209
|
-
| `0.1.5-rc.1` | 目录 / Directory | 通过 / Pass |
|
|
210
|
-
| `0.1.5-rc.2` | 目录 / Directory | 通过 / Pass |
|
|
211
|
-
| `0.1.5-rc.3` | 目录 / Directory | 通过 / Pass |
|
|
212
|
-
| `0.1.6-alpha.1` | 目录 / Directory | 通过 / Pass |
|
|
213
|
-
| `0.1.6-alpha.2` | 目录 / Directory | 通过 / Pass |
|
|
214
|
-
| `0.1.7-alpha.1` | 声明式 / Declarative | 通过 / Pass |
|
|
215
|
-
| `0.1.7-alpha.2` | 声明式 / Declarative | 通过 / Pass |
|
|
216
|
-
| `0.1.7-rc.1` | 声明式 / Declarative | 通过 / Pass |
|
|
217
|
-
| `0.1.7-rc.2` | 声明式 / Declarative | 通过 / Pass |
|
|
218
|
-
| `0.2.0-rc.1` | 声明式 / Declarative | 通过 / Pass |
|
|
219
|
-
| `0.2.0-rc.2` | 声明式 / Declarative | 通过 / Pass |
|
|
220
|
-
| `0.2.1-alpha.1` | 声明式 / Declarative | 通过 / Pass |
|
|
221
|
-
| `0.2.1-alpha.2` | 声明式 / Declarative | 通过 / Pass |
|
|
222
|
-
|
|
223
|
-
原生 PTC 另在最低版、稳定最新版和预发布最新版(`0.1.5-alpha.1`、`0.2.0-rc.2`、`0.2.1-alpha.2`)验证独立读取并行、写入顺序、权限拒绝与部分失败保留。0.1.7 的两个 alpha 版没有预设文档读取接口,因此归属核验要求唯一、启用的 Loader 声明;其他声明式版本核验实际活跃组合的来源标记。桌面宿主使用自己的安装位置解析管理 API,避免加载第二份宿主模块。
|
|
224
|
-
|
|
225
|
-
Native PTC additionally passed parallel independent reads, ordered writes, permission denial and partial-failure preservation on the minimum, stable latest and prerelease latest versions. The two 0.1.7 alpha hosts lack the document reader, so ownership requires a unique enabled Loader declaration; later hosts verify the marker in the active composition. Desktop management resolves APIs from the running host installation to avoid loading a separate host module.
|
|
226
|
-
|
|
227
|
-
官方 macOS arm64 的 `0.1.7-rc.1`、`0.1.7-rc.2`、`0.2.0-rc.1`、`0.2.0-rc.2` 发行包均在隔离应用副本中启动,验证原生后端的预设安装、改名重装、语言和版本状态。副本保留原始 CFBundleName,仅隔离应用身份、profile 与测试端口。Safari 最新 alpha.2 与官方 Desktop rc.2 的实际界面验收另留截图。DSH NEXT 2.0.17-next(Harness rc.2)也使用隔离副本验收。
|
|
15
|
+
| 官方 Web | `0.1.5-alpha.1`、`alpha.2`、`rc.1`、`rc.2`、`rc.3` | 目录 |
|
|
16
|
+
| 官方 Web | `0.1.6-alpha.1`、`alpha.2` | 目录 |
|
|
17
|
+
| 官方 Web | `0.1.7-alpha.1`、`alpha.2`、`rc.1`、`rc.2` | 声明式 |
|
|
18
|
+
| 官方 Web | `0.2.0-rc.1`、`rc.2` | 声明式 |
|
|
19
|
+
| 官方 Web | `0.2.1-alpha.1` ~ `alpha.5`、`rc.1`、`rc.2` | 声明式 |
|
|
20
|
+
| 官方 Desktop(macOS arm64) | `0.1.7-rc.1`、`0.1.7-rc.2`、`0.2.0-rc.1`、`0.2.0-rc.2`、`0.2.1-rc.2` | 声明式 |
|
|
21
|
+
| DSH NEXT | `2.0.17-next`(内置 Harness `0.2.0-rc.2`) | 声明式 |
|
|
22
|
+
|
|
23
|
+
0.3.8 的增量功能另在 NEXT `2.0.19-next`(内置 Harness `0.2.1-rc.2`)上验证。
|
|
24
|
+
|
|
25
|
+
## 验证方法
|
|
26
|
+
|
|
27
|
+
- 每个宿主使用独立 profile,宿主依赖固定版本并经过审计,不影响日常使用的应用和配置。
|
|
28
|
+
- 模型响应由本地确定性程序提供,经过真实的宿主 Agent 和工具链。**未调用外部模型**,因此不属于效果评测。
|
|
29
|
+
- Web 端使用 Chromium;历史原生验收在 macOS arm64 上运行,rc.4 官方 x64 包另通过 Rosetta 验证。Windows Bash 另使用 Windows 11 ARM64 VM 运行真实 x64 桌面发行包,未使用 Intel 实机。
|
|
30
|
+
- 每次发布均同时运行全量单元测试、类型检查、构建和客户端语法检查。
|
|
31
|
+
|
|
32
|
+
## 未覆盖的范围
|
|
228
33
|
|
|
229
|
-
|
|
34
|
+
- Linux 原生客户端、Intel 实机、Safari 专属行为;Windows 桌面仅覆盖下述内置 Bash 专项,未逐版重跑全部插件功能
|
|
35
|
+
- 真实模型服务、完整的升级与自定义配置回归、所有工具组合
|
|
36
|
+
- 窄窗口:宽度为 440px 时,宿主侧边栏仅为插件保留 156px,视为不可用(未安装插件时设置页存在同样问题)
|
|
37
|
+
- 原生目录选择器的确认与取消(系统辅助功能接口超时)
|
|
38
|
+
- MCP 客户端能否根据连接 Prompt 自动完成配置
|
|
39
|
+
- 各版本的冷下载安装(测试复用了缓存的宿主依赖)
|
|
230
40
|
|
|
231
|
-
|
|
41
|
+
依赖声明中的版本范围不等同于验证结论,未列出的版本需单独验证。
|
|
232
42
|
|
|
233
|
-
|
|
43
|
+
## 安装问题排查
|
|
234
44
|
|
|
235
|
-
|
|
45
|
+
预发布版本的 `^` 范围可能解析到其他 alpha/rc 版本,导致宿主组件版本混用。例如:
|
|
236
46
|
|
|
237
|
-
|
|
47
|
+
- 安装 `0.1.6-alpha.1` 时混入 alpha.2,插件加载前即报 `watchUserPatches` 导出缺失。
|
|
48
|
+
- 安装 `0.1.7-alpha.1` 时混入 rc.1。
|
|
238
49
|
|
|
239
|
-
|
|
50
|
+
测试环境通过精确的 overrides 解决该问题,但无法修复上游的依赖声明。遇到类似错误时,请先核对实际安装的宿主组件版本。
|
|
240
51
|
|
|
241
|
-
|
|
52
|
+
按照 npm 规则,`>=0.1.5-alpha.1 <0.3.0` 这类写法会排除 `0.1.6-alpha.1`、`0.2.0-rc.2` 等预发布版本,因此 peer 依赖只能按系列分别列出。
|
|
242
53
|
|
|
243
|
-
|
|
54
|
+
`zod`、`js-yaml`、`@modelcontextprotocol/sdk` 为运行依赖,不打包进插件。Windows Shell 另随包携带独立进程运行的第三方程序,其来源、许可证和对应源码见[运行时说明](../runtime/windows/README.md)。
|
|
244
55
|
|
|
245
|
-
|
|
56
|
+
## 0.3.10 · 2026-11-03
|
|
246
57
|
|
|
247
|
-
|
|
58
|
+
以下为 2026-11-02 ~ 03 各阶段的 Windows Bash 验证,保留各批次的代码基线与验收范围,不改变下方历史版本记录。
|
|
248
59
|
|
|
249
|
-
|
|
60
|
+
- **发布检查**:0.3.10 的 326 项全量单测、类型检查、构建及客户端语法检查通过。发布准备仅更新版本元数据与文档,执行代码沿用上述已验收实现;未为版本号变化重复全部宿主矩阵。
|
|
61
|
+
- **官方预设命令环境配置**:新增独立的「同时应用到官方预设」开关,默认关闭;开启后标准、PTC、创建模式使用内置 Bash,极简模式保留持久终端,自定义预设不自动修改。设置自动保存,与预算、删除保护范围互不影响;总开关关闭后保留适用范围选择。运行中的 Agent 固定选择,子 Agent 按实际运行时归属继承;准备失败的父子 Agent 不会被其他会话的恢复误激活。325 项单测、类型检查、构建、客户端语法通过,24 个版本各 15 项 Shell 组件回归,共 360 项通过;组件测试不等同于完整 Windows 宿主验收。
|
|
62
|
+
- **官方预设专项实测**:Windows Web 的 `0.1.5-alpha.1`、`0.1.6-alpha.1`、`0.1.7-rc.2`、`0.2.0-rc.2`、`0.2.1-alpha.3`、`0.2.1-rc.5` 各 13 组,共 78 组通过。随后补充父准备失败继承修复,最终运行包在最低版与 rc.5 Web 各 13 组、官方 rc.5 Desktop 与 NEXT `2.0.21-next`(内置 rc.4)各 13 组复验通过。覆盖 Standard/PTC/Create 真实 Bash、默认关闭时的 PowerShell、Minimal 持久终端、双向自动保存、运行中选择保留、限权拒绝与宿主单次审批、中文空格路径及文件副作用、完整进程冷重启。最终包未将其余四个 Web 版本或历史桌面完整套件全部重跑;原 155/558 组记录使用新增官方范围开关前的包。
|
|
63
|
+
- **命令环境界面补充**:只有检测为 Windows 时可编辑;其他平台的两个开关均灰显,顶部悬浮提示「仅 Windows 生效」。Bash 默认开启、官方预设范围默认关闭,保留已有选择;就绪状态与设置文案对齐,长版本号可换行。补充后 326 项单测、客户端语法通过;32 个官方静态前端目标(24 Web、8 Desktop/NEXT)共 192 组浏览器检查通过,覆盖双语提示、禁用时无写入、默认值、原有按钮气泡与窄窗口布局。该界面回归使用夹具服务,不属于重新运行完整 Windows 宿主验收。
|
|
64
|
+
- **气泡主题**:所有共享提示气泡跟随宿主主题,浅色白底深色文字、深色深底浅色文字,箭头、边框和阴影同步。最低版、rc.5、alpha.2 与 NEXT 的静态前端共 32 组检查通过,覆盖按钮、禁用开关、模态对话框和气泡显示中切换主题;未重跑完整宿主。
|
|
65
|
+
- **完整 Windows Web**:`0.2.1-rc.4`、`rc.5` 的真实 Loader、预设、Agent、PTC 和 Windows ACL 沙箱已验收。`run_code → bash` 通过真实 DSH 界面的批准、拒绝、取消,以及禁用审批、下一次再次拒绝和后台前置拒绝;批准命令实际退出 0 且只写入一次,其他 6 个未批准路径无文件副作用,会话仍为 workspace-write。已核对两版实际加载的关键组件版本。
|
|
66
|
+
- **最低版本 Windows Web 补测**:`0.1.5-alpha.1`、`alpha.2`、`rc.1`、`rc.2`、`rc.3` 各通过 31 组完整进程检查,共 155 组,覆盖预设、设置自动保存、旧版 codeRuntime、Bash、真实界面审批、删除保护、非零退出、取消、实际文件副作用和 CLI 冷重启。旧版 120ms 超时会终止命令,不转后台;超过延迟窗口后无取消或超时写入。各版实际 profile 中可解析的 229/230 个 DSH 组件均为目标版本,12 个执行关键组件逐个核对;使用官方 Node `24.18.1` ARM64 与原有 Cordis `4.0.2` 组合。
|
|
67
|
+
- **中间 16 版 Windows Web 补齐**:`0.1.6-alpha.1/alpha.2`、`0.1.7-alpha.1/alpha.2/rc.1/rc.2`、`0.2.0-rc.1/rc.2`、`0.2.1-alpha.1` ~ `alpha.5` 及 `rc.1` ~ `rc.3` 全部完成真实 Windows Web 宿主验收。0.1.6 两版各 34 组,其余 14 版各 35 组,共 558 组通过,覆盖预设可用、设置自动保存、PTC → Bash、真实审批、三档删除保护、非零退出、超时、取消、实际文件和完整 CLI 冷重启。关闭 Bash 后,运行中 Agent 保留选择,新 Agent 能使用官方 PowerShell。
|
|
68
|
+
- **中间版本审计与超时差异**:各版在 Windows 11 ARM64 VM 使用官方 Node `24.18.1` ARM64、按各版官方声明固定的 Cordis/Schemastery 依赖和独立 profile;实际安装的 236 ~ 296 个 DSH 组件均无混版,18 个关键模块逐个核对,加载的开发包运行代码哈希一致。0.1.6 的 120ms 超时会终止命令,其余版本会转为一次后台任务并正常完成;从 `0.2.1-alpha.3` 起另核对 v4 的任务注册、完成事件。取消检查先确认真实 Bash 进程及启动文件,再取消并等待超过 12 秒,确认进程退出且无延迟写入。这是 Windows Web 验收,历史原生桌面范围仍按下述记录单列。
|
|
69
|
+
- **旧版模块定位修复**:修复无 `profileContext` 的目录预设在 npm 嵌套依赖布局下找不到宿主 Shell 模块、无法创建会话的问题;改从实际 root profile 解析,保留新版显式安装锚点的优先级。新增回归先复现再修复,全量 317 项单元测试与 24 版的 216 项 Shell 回归通过。上述 0.1.5 五版及中间 16 版均已使用修复包完成 Windows Web 验收。
|
|
70
|
+
- **0.1.5 桌面边界**:查过的五个 Windows 安装包地址及对应 `dsh-desktop` npm 元数据均返回 404,未完成原生旧桌面验收;完整 Web 结果不能代替桌面结果。下方完整 Desktop/NEXT、删除保护与超时记录使用模块定位修复前的包;本轮最终包的官方预设专项另见上文,不将两批结果合称为完整套件复验。
|
|
71
|
+
- **完整 Windows Desktop / NEXT**:官方 Desktop `0.2.1-rc.5` 与 NEXT `2.0.21-next` 串行运行真实发行安装包;后者内置 Harness `0.2.1-rc.4`。两端实际 Host 均为 x64、Node `24.21.0`、Electron `44.7.0`,11 个关键宿主包已核对。通过原生插件页安装和启用当时的开发 tgz、预设可用、Bash 开关双向自动保存、预算保留、跨标签恢复及完整应用冷重启;测试期间无页面脚本错误。
|
|
72
|
+
- **桌面执行与审批**:两端真实 `run_code → bash` 均通过数组、中文带空格文件名、GNU grep、BusyBox、非零退出、运行取消及超时转后台并完成。受限调用明确拒绝,宿主单次批准只执行一次;拒绝、取消、never、后续普通调用和后台前置拒绝均无文件副作用,会话保持 workspace-write。关闭 Bash 后旧 Agent 保留选择,新 Agent 使用宿主 PowerShell。
|
|
73
|
+
- **桌面删除保护**:两端通过严格模式下 MSYS `cmd.exe //c` 删除的拒绝、取消和批准;在 Workspace write + ask 下,删除确认与该条 Bash 的沙箱外权限申请分别批准,才删除任务临时文件。拒绝或取消均保留文件,宽松和关闭模式不增加插件确认。宿主 Full access 默认同时设置 approval=never,此时严格删除申请被自动拒绝,插件不改变宿主策略。
|
|
74
|
+
- **桌面安装来源**:官方包来自 [DeepSeek Windows nightly](https://download.deepseek.com/dsh-desk/feeds/win-x64/nightly.yml),SHA-512 与更新源一致,Windows Authenticode 有效。NEXT 来自 [NEXT Windows 下载接口](https://www.dshdesktop.cn/api/downloads/windows),携带 next 频道与目标版本请求头;安装程序未签名且接口未提供公开摘要,仅核对 HTTPS 来源、产品版本及传输前后 SHA-256,不能视为签名验证通过。
|
|
75
|
+
- **限权边界**:MSYS 无法在官方 Windows 限权沙箱中创建内部命名管道。插件保留 Bash,在进程启动前明确拒绝受限调用;仅接受宿主批准的单次 `danger-full-access` 调用策略,之后的命令仍需审批,不修改会话权限或切换 Shell。另在显式完全访问的测试会话中验证了数组、中文文件读写和 GNU grep。
|
|
76
|
+
- **测试环境对照**:前期 Node 22.23.4 ARM64 出现不加载插件也能复现的 fs.mkdir/ACL 锁文件 EPERM。保留失败记录后,在同一 VM 用官方 Node 24.18.1 完成上述完整审批与设置页验收;没有修改系统 ACL 或关闭防护。
|
|
77
|
+
- **实机范围**:专用 Windows 11 ARM64 VM;两套载荷各通过 7 项执行检查。桌面发行包和 Bash/MSYS/grep 均通过系统 x64 兼容执行,不能视为原生 x64 硬件或原生 ARM64 Bash 验证。完整桌面专项仅覆盖上述两款发行包,未重跑所有历史 Windows 桌面版本或外部模型评测。
|
|
78
|
+
- **真实工具组件**:rc.4 的 12 项通过,覆盖根/子作用域过滤、按会话保留 Shell 选择、PowerShell 切回、单次副作用、超时/取消、删除保护及单次审批、宿主拒绝、后台任务和销毁隔离。这组使用真实工具和 Windows 进程,未证明完整限权沙箱可运行 Bash。
|
|
79
|
+
- **固定命令题集**:100 项中 97 项通过,原 45 项全部通过。剩余 3 项涉及 `.cmd` 引号及 MSYS 对原生 `/` 参数的转换。题目与预期结果未修改;不全局关闭路径转换,不失败重放。
|
|
80
|
+
- **逐版接口**:从 `0.1.5-alpha.1` 到 `0.2.2-alpha.2` 的全部 24 个 npm 发布版本逐版通过,共 591 项 Shell 作用域、删除保护和设置检查。旧版使用真实 SettingsProvider;新版使用真实 Loader 与 SettingsForms,ConfigEditor 持久化为临时文件夹具。alpha 使用真实 native executor、batch API 和会话投影,删除保护整组通过。24 版真实 provider 另在 macOS 通过 240 项执行检查,其中沙箱拒绝为桩;不等同于 24 个完整 Windows 宿主验收。
|
|
81
|
+
- **逐版 Windows 执行**:上述 24 版均在 Windows VM 中独立安装并核对真实 provider 依赖图,共 288 项通过,覆盖实际 Bash 进程、中文与 stdin/stderr、非零退出、超时/取消、单次副作用、截断与完整溢出文件、工具隔离、限权前置拒绝及环境变量屏蔽。每版使用独立 Node 进程;批准策略检查使用组件服务,不冒充完整 ACL 宿主审批。运行时自身 13 项损坏、并发、硬链接/符号链接和环境边界测试也在 Windows 通过。
|
|
82
|
+
- **安装范围**:查询时 npm latest/next 为 `0.2.1-rc.5`,alpha 为 `0.2.2-alpha.2`。24 版的 360 项 peer 范围检查通过;已补齐新 alpha 系列及其 Cordis/Schemastery 预发布范围。开发包在独立 alpha 依赖图中真实安装,并通过包根插件的 Loader、SettingsForms 激活与设置读写;该检查的 ConfigEditor 为夹具。
|
|
83
|
+
- **界面与单元**:32 个官方静态前端目标(24 Web、8 Desktop/NEXT)通过 128 项设置开关、自动保存、错误与非 Windows 禁用检查,无页面脚本错误。RPC 为夹具,不代表运行中桌面验收。316 项单元测试、类型检查、构建和客户端语法检查通过。完整 Windows Web 设置页另通过双向自动保存、预算与平台状态保留、跨标签和冷重启检查。
|
|
84
|
+
- **并发复查**:另在全部 24 版通过 192 项 Shell 检查,新增覆盖准备中的父子 Agent、重复生命周期事件、父子失败后兄弟恢复,以及真实离线准备服务的 Promise 共享、失败重试和状态快照隔离;这轮在 macOS 运行,未启动 Windows 二进制或完整宿主。
|
|
85
|
+
- **环境与删除识别复查**:修复 `PATH`/`Path` 键序可能清掉最终 PATH,以及别名覆盖固定 BusyBox 选项的问题;24 版真实环境合并函数通过 240 项检查,Windows VM 另通过 168 项实际进程检查。严格删除保护补充识别 MSYS 的 `cmd.exe //c`、`//k` 写法,全部 24 版通过 360 项删除保护组件回归;普通输出中的删除命令文字不受影响。这些检查不代表完整宿主或外部模型验收。
|
|
86
|
+
- **检视修复**:准备失败的 Agent 不再暴露另一 Shell,也不会因兄弟 Agent 恢复而重新获得执行能力;非标准 provider 关闭后,子 Agent 保留父会话选择。损坏的清单在准备前拒绝;Windows PATH 别名遵循 Node 优先级,启动脚本与导出函数环境变量被遮蔽。
|
|
87
|
+
- **体积与模型**:载荷压缩资产约 4.50 MB;最近开发包约 9.1 MB,包含 BusyBox 对应源码。以上均使用固定测试命令或本地确定性模型响应,没有调用外部模型,不属于编码效果评测。
|
|
250
88
|
|
|
251
|
-
|
|
89
|
+
## 0.3.9 · 2026-10-27 ~ 28
|
|
252
90
|
|
|
253
|
-
|
|
91
|
+
本节区分初次发布的界面验收和 npm 发布前的审批修复复验,不代表上方 26 个完整宿主均重新验收。
|
|
254
92
|
|
|
255
|
-
|
|
93
|
+
- **初次界面验收**:整合候选标记为 0.3.8,首次 0.3.9 发布的运行代码与该候选逐字节一致;284 项测试通过。使用官方 React、主题和控件配合本地模拟服务,30 个前端目标通过 390 组交互(Web `0.1.5-alpha.1` ~ `0.2.1-rc.4`、6 个官方桌面前端、2 个 NEXT 前端),覆盖中英文、深浅主题、自动保存与失败、预设状态、自定义下拉、窄布局、运行圆圈和长列表。**前端检查不等于真实宿主验收。**
|
|
94
|
+
- **真实界面与持久化**:官方 Desktop `0.2.1-rc.3`、NEXT `2.0.19-next`,以及 10 月 28 日追加的完整 Web `0.2.1-rc.4` 和官方 Desktop `0.2.1-rc.4` x64/Rosetta 宿主,通过设置、真实 MCP SDK 连接、工作区刷新与授权撤销、旧版记忆增删改与取消、任务投影与虚拟滚动、冷重启恢复;无页面脚本错误。rc.4 使用固定并审计的 274 个官方组件。
|
|
95
|
+
- **本次修复回归**:修复 `run_code` 子工具审批无法转交调用方的问题;全量 289 项测试、类型检查、构建与客户端语法检查通过。20 个原有 Web 引擎版本(最低 `0.1.5-alpha.1` 至 `0.2.1-rc.2`)和 `0.2.1-rc.4` 各通过 28 项审批/删除保护回归。
|
|
96
|
+
- **真实审批链路**:修复包在 Web `0.1.5-alpha.1`、Web `0.2.1-rc.4`、官方 Desktop `0.2.1-rc.3` / `0.2.1-rc.4`(x64/Rosetta)、NEXT `2.0.19-next` 串行通过 `run_code → bash` 的准确参数转交、批准、拒绝、取消、跨任务隔离、过期决策拒绝、关闭开关留给 DSH 和容量恢复;批准仅删除任务专用临时文件,拒绝与取消均保留文件。
|
|
97
|
+
- **rc.4 MCP 全流程**:官方 Web 和官方桌面 x64/Rosetta 均另复验预设改名重装与备份、旧预算默认值与输入/冲突校验、模型与思考强度实际传入、显式复用与后续消息、公开历史分页、目录同意/拒绝、3 连接和 3 任务上限、取消回收、子 Agent 历史及断线重连。
|
|
98
|
+
- 所有真实宿主使用本地确定性模型响应,**未调用外部模型**。侧栏最多一项展开,默认任务,也可全部收起;只包含旧版记忆,不含新版记忆、类型切换或迁移。
|
|
99
|
+
- rc.4 桌面使用官方 mac-x64 nightly ZIP(SHA-512 与官方清单一致,原应用签名校验通过),在 macOS arm64/Rosetta 下运行;不是改版号的 rc.3 或源码替代包。mac-arm64 频道仍为 rc.3,rc.4 arm64 ZIP 返回 404;**原生 arm64 rc.4 包尚未验收**。未重新验证原生目录选择器。
|
|
256
100
|
|
|
257
|
-
|
|
258
|
-
|---|---|---|---|
|
|
259
|
-
| `0.1.5-alpha.1` | 目录 / Directory | 通过 / Pass | 通过 / Pass |
|
|
260
|
-
| `0.1.5-rc.3` | 目录 / Directory | 通过 / Pass | 未重测 / Not rerun |
|
|
261
|
-
| `0.1.6-alpha.1` | 目录 / Directory | 通过 / Pass | 未重测 / Not rerun |
|
|
262
|
-
| `0.1.6-alpha.2` | 目录 / Directory | 通过 / Pass | 未重测 / Not rerun |
|
|
263
|
-
| `0.1.7-alpha.1` | 声明式 / Declarative | 通过 / Pass | 未重测 / Not rerun |
|
|
264
|
-
| `0.1.7-alpha.2` | 声明式 / Declarative | 通过 / Pass | 未重测 / Not rerun |
|
|
265
|
-
| `0.1.7-rc.1` | 声明式 / Declarative | 通过 / Pass | 未重测 / Not rerun |
|
|
266
|
-
| `0.1.7-rc.2` | 声明式 / Declarative | 通过 / Pass | 未重测 / Not rerun |
|
|
267
|
-
| `0.2.0-rc.1` | 声明式 / Declarative | 通过 / Pass | 未重测 / Not rerun |
|
|
268
|
-
| `0.2.0-rc.2` | 声明式 / Declarative | 通过 / Pass | 通过 / Pass |
|
|
269
|
-
| `0.2.1-alpha.1` | 声明式 / Declarative | 通过 / Pass | 通过 / Pass |
|
|
101
|
+
## 各版本验证记录
|
|
270
102
|
|
|
271
|
-
|
|
103
|
+
### 0.3.8 · 2026-10-26
|
|
272
104
|
|
|
273
|
-
|
|
105
|
+
- **基础功能**:26 个目标通过 MCP 连接验收,包括设置保存、真实 MCP SDK 客户端认证、原生 Super Code 会话、重复请求去重、补充要求和运行中取消。
|
|
106
|
+
- **设置页与目录审批**:在最低版本 Web `0.1.5-alpha.1`、最高版本 Web `0.2.1-rc.2`、Desktop `0.2.1-rc.2` 和 NEXT `2.0.19-next` 上逐个串行验证:首页、目录勾选、自动保存、凭证展示、设置页关闭时的全局目录审批(批准后可在新目录执行,拒绝则不授权),以及连接页的会话列表和断开操作。页面脚本错误为零。
|
|
107
|
+
- **工作区目录修复**:修复 Workspace 路径 getter 读取失败、有效目录被判定为「不可用」导致无法启用 MCP 的问题。上述四端均使用含中文和空格的真实工作区复验。
|
|
108
|
+
- **会话列表与显式复用**:253 项测试通过。仅在 NEXT `2.0.19-next` 上进行增量验证(握手发现六个工具,查询到四个目录下的六条历史任务,用户已保存的上限 10 得到保留),未重新运行 26 个目标。
|
|
109
|
+
- **重启**:冷重启后保留开关、凭证和任务 ID;未完成的任务不会自动重发,也不会误报成功。
|
|
110
|
+
- 每条对话的在线连接数仅表示当前关联的连接(0 或 1),并非多客户端租约或会话锁。
|
|
274
111
|
|
|
275
|
-
|
|
112
|
+
### 0.3.7 · 2026-10-24
|
|
276
113
|
|
|
277
|
-
|
|
114
|
+
- 229 项测试通过。
|
|
115
|
+
- 26 个目标各通过 25 项删除保护和任务状态引擎检查(旧设置、官方预设隔离、逐次确认、新旧日志、嵌套 PTC、文件归属,以及任务状态不影响模型请求)。
|
|
116
|
+
- 最低/最高版本 Web、Desktop `0.2.1-rc.2`、NEXT `2.0.17-next` 另通过 7 项任务状态页面检查(含 373/374 行长列表的滚动和键盘访问)和 6 项删除保护设置检查。
|
|
117
|
+
- 严格模式的确认由测试程序应答宿主审批服务,不等同于人工点击原生确认框。
|
|
278
118
|
|
|
279
|
-
|
|
119
|
+
### 0.3.6 · 2026-10-20
|
|
280
120
|
|
|
281
|
-
|
|
121
|
+
- 26 个目标各通过 6 项页面/真实 Agent 检查和 1 项冷重启检查;195 项测试通过。
|
|
122
|
+
- 旧宿主没有 `ptcRuntime` 时,不会写入不支持的时限字段。
|
|
123
|
+
- 测试候选版本标记为 `0.3.5`;发布的 `0.3.6` 仅修改了包版本和预设版本标记,其余运行文件已逐个核对一致。
|
|
282
124
|
|
|
283
|
-
|
|
284
|
-
CLI identity and startup, plugin installation and activation, configuration loading, Web startup, and persistent memory reads after restart.
|
|
285
|
-
- Super Code 预设发现及状态、工作区和会话创建、选模、发送提示词,以及 `super_code_method` 返回非错误结果。
|
|
286
|
-
Super Code preset discovery and status, workspace/session creation, model selection, prompting, memory persistence/restart reads, and a non-error `super_code_method` result.
|
|
287
|
-
- 宿主 npm 与 profile pnpm 锁文件审计。插件不再依赖新版宿主已停用的目录预设辅助包;旧版目录副本仍由插件管理。
|
|
288
|
-
Host npm and profile pnpm lockfile audits. The plugin no longer depends on the retired directory-preset helper package; it still manages directory copies on older hosts.
|
|
125
|
+
### 运行预算 · 2026-10-19
|
|
289
126
|
|
|
290
|
-
|
|
127
|
+
- 26 个目标各通过 11 项页面检查、6 项真实会话检查和 3 项冷启动检查;192 项测试通过。
|
|
128
|
+
- 20 个 Web 版本在 800px 宽窗口下通过控件边界、滚动和点击命中检查。
|
|
129
|
+
- 15 个 Web 版本和 NEXT 使用真实 Node PTC 验证了默认时限和显式指定的更短时限;`0.1.5` 的五个版本没有 `ptcRuntime`,时限字段正确禁用。
|
|
130
|
+
- 修复语言偏好尚未保存时插件即重载、导致语言回退的问题。
|
|
291
131
|
|
|
292
|
-
|
|
132
|
+
### 0.3.5 · 2026-10-18
|
|
293
133
|
|
|
294
|
-
|
|
134
|
+
- 新增 `0.2.1-alpha.3`、`alpha.4`、`alpha.5`、`rc.1`、`rc.2`,五个版本均通过 Web/浏览器和原生 Skill 加载检查。
|
|
135
|
+
- Desktop `0.2.1-rc.2` 从官方地址下载并校验 SHA-512,通过原生插件页安装和完整界面检查。
|
|
136
|
+
- 修复 v4 日志迁移后旧任务记录不显示的问题:v4 将来源从 `{kind: "plugin", plugin: "…"}` 改为 `{kind: "plugin:…"}`,旧读取器会跳过此类记录。现已兼容两种格式,任务投影版本从 1 升至 2,使旧缓存重新回放。已使用官方迁移器和真实 v3 会话验证。
|
|
137
|
+
- 173 项测试通过。
|
|
295
138
|
|
|
296
|
-
|
|
139
|
+
### 0.3.4 · 2026-10-15
|
|
297
140
|
|
|
298
|
-
|
|
141
|
+
仅修改文档和评测快照。构建后的 112 个运行文件与 0.3.3 逐字节一致,沿用 0.3.3 的验证结果。171 项测试通过。
|
|
299
142
|
|
|
300
|
-
|
|
143
|
+
### 0.3.3 · 2026-10-14
|
|
301
144
|
|
|
302
|
-
|
|
145
|
+
- Web `0.1.5-alpha.1`、`0.2.0-rc.2`、`0.2.1-alpha.2` 通过安装、主/子 Agent、记忆、重启和浏览器检查(统一侧边栏、记忆维护、动画、320/400/560px 宽度和缩放)。
|
|
146
|
+
- 前端 Skill 在上述三个版本上通过六条原生加载路径。
|
|
147
|
+
- Desktop `0.2.0-rc.2` 从 0.3.2 升级至 0.3.3:旧预设保留旧版本并提示差异,确认重装后完成更新。
|
|
148
|
+
- 171 项测试通过。
|
|
303
149
|
|
|
304
|
-
|
|
150
|
+
### 0.3.2 · 2026-10-11
|
|
305
151
|
|
|
306
|
-
|
|
152
|
+
- 上述三个 Web 版本和 Desktop `0.2.0-rc.2` 通过界面改版检查。
|
|
153
|
+
- 启动时等待语言偏好的修复仅在三个 Web 版本和单元测试中复验,原生窗口截图因控制接口超时未能获取。
|
|
154
|
+
- 158 项测试通过。
|
|
307
155
|
|
|
308
|
-
|
|
156
|
+
### 0.3.1 · 2026-10-10
|
|
309
157
|
|
|
310
|
-
|
|
158
|
+
- 15 个 Web 版本(`0.1.5-alpha.1` ~ `0.2.1-alpha.2`)全部通过。
|
|
159
|
+
- Desktop `0.1.7-rc.1`、`rc.2`、`0.2.0-rc.1`、`rc.2` 和 NEXT `2.0.17-next` 通过原生后端的预设安装、改名重装、语言和版本检查。
|
|
160
|
+
- 原生 PTC 在 `0.1.5-alpha.1`、`0.2.0-rc.2`、`0.2.1-alpha.2` 上验证了并行读取、写入顺序、权限拒绝和部分失败。
|
|
161
|
+
- `0.1.7` 的两个 alpha 版本没有预设文档读取接口,归属核验改为要求存在唯一且已启用的 Loader 声明。
|
|
162
|
+
- 更早版本和 `0.2.1-alpha.1/alpha.2` 的 Desktop 安装包下载返回 404,原生窗口未验证。
|
|
311
163
|
|
|
312
|
-
|
|
164
|
+
### 0.3.0 · 2026-10-09
|
|
313
165
|
|
|
314
|
-
|
|
166
|
+
`0.1.5-alpha.1`、`0.2.0-rc.2`、`0.2.1-alpha.1` 通过原生 PTC、主/子 Agent 记忆、重启读取和 Playwright 网页检查。中间版本未重新测试。
|
|
315
167
|
|
|
316
|
-
|
|
168
|
+
### 0.2.1 · 2026-10-08
|
|
317
169
|
|
|
318
|
-
|
|
170
|
+
- `0.2.0-rc.2`、`0.2.1-alpha.1` 通过验证,包括 Safari 检查和插件停用/重新启用。
|
|
171
|
+
- 自 `0.2.1-alpha.1` 起,宿主通过 `<插件>/locale/*.json` 读取插件名称和描述,子路径不再读取独立的 `package.json`。插件已为根路径、`/dsh`、`/super-code` 导出中英文资源。
|
|
172
|
+
- 宿主拆分了 composer 的 `stats` 并移除了 `invariant` 导出,插件未使用这些接口,无需适配。
|
|
173
|
+
- `0.2.1-alpha.1` 没有 Desktop 安装包,原生端未验证。
|
|
319
174
|
|
|
320
|
-
|
|
175
|
+
### 0.2.0 · 2026-09-30
|
|
321
176
|
|
|
322
|
-
|
|
177
|
+
`0.1.5-alpha.1` ~ `0.2.0-rc.2` 的十个版本通过 Web/CLI 和浏览器检查。Desktop `0.2.0-rc.2` 通过原生窗口检查,包括中英文切换和重启后恢复。
|
|
323
178
|
|
|
324
|
-
|
|
179
|
+
### 0.1.4
|
|
325
180
|
|
|
326
|
-
|
|
181
|
+
`0.1.5-rc.3`、`0.1.6-alpha.1`、`alpha.2`、`0.1.7-alpha.1`、`alpha.2`、`rc.1`、`rc.2` 通过 Web/CLI 冒烟检查。
|