dsh-github-copilot 0.4.0-alpha.82 → 0.4.0-alpha.83
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/CHANGELOG.md +6 -0
- package/README.md +10 -8
- package/README.zh.md +10 -8
- package/deployment-baseline.json +5 -1
- package/docs/auto-task-routing.md +24 -1
- package/docs/automatic-model-routing.md +8 -3
- package/lib/client.js +2 -2
- package/lib/client.js.map +1 -1
- package/lib/index.js +66 -20
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,11 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## 0.4.0-alpha.83 (prepared)
|
|
4
|
+
|
|
5
|
+
- Reduce avoidable Auto uncertainty through bounded difficult-task continuation and isolated fenced mechanical-transform recognition (#262); preserve unknown for new, incomplete and ambiguous work.
|
|
6
|
+
- Prioritize current and preceding user requests in the bounded semantic projection, preserving complete JSON rows, chronological order and omitted-context downshift protection.
|
|
7
|
+
- Call the optional semantic experiment only for locally unknown demand. Explain local rule coverage separately from absent conversation context; retain off-by-default semantics, category policy and native turn ownership.
|
|
8
|
+
|
|
3
9
|
## 0.4.0-alpha.82 (prepared)
|
|
4
10
|
|
|
5
11
|
- Explain missing native Turn Usage on completed Copilot replies through a plugin-owned read-only conversation projection and a compact, keyboard-accessible footer disclosure (#259).
|
package/README.md
CHANGED
|
@@ -109,7 +109,7 @@ The Host reads GitHub's internal quota endpoint using the existing canonical OAu
|
|
|
109
109
|
|
|
110
110
|
## Install and sign in
|
|
111
111
|
|
|
112
|
-
The commands below target the package version `0.4.0-alpha.
|
|
112
|
+
The commands below target the package version `0.4.0-alpha.83`. Versioned URLs describe the intended release artifacts, not proof that publication or local activation has completed; use them only once that Release and its checksums are available. Install into the profile you use (replace `web` when targeting another profile):
|
|
113
113
|
|
|
114
114
|
Before installing/updating, unpack the **checksum-verified** archive into a temporary directory and run its read-only composition preflight (replace all paths with absolute paths for the intended profile):
|
|
115
115
|
|
|
@@ -122,10 +122,10 @@ Include any launcher patch files with repeated `--patch /absolute/file` argument
|
|
|
122
122
|
For a standalone named profile only, the supported CLI command is:
|
|
123
123
|
|
|
124
124
|
```sh
|
|
125
|
-
dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.
|
|
125
|
+
dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.83/dsh-github-copilot-0.4.0-alpha.83.tgz
|
|
126
126
|
```
|
|
127
127
|
|
|
128
|
-
For Desktop profiles, use Desktop's native package manager with `dsh-github-copilot@0.4.0-alpha.
|
|
128
|
+
For Desktop profiles, use Desktop's native package manager with `dsh-github-copilot@0.4.0-alpha.83` after the same preflight and only when registry access is permitted. Desktop reserves its `desktop` profile from CLI management. If registry access is blocked or unavailable, stop; the documented [controlled offline CLI procedure](./docs/npm-distribution.md#controlled-offline-cli-maintenance-for-standalone-profiles) is not a Desktop workaround.
|
|
129
129
|
|
|
130
130
|
Then open the Models UI listed above, find **GitHub Copilot**, select **Sign in**, and complete the GitHub device-code flow. Plugin installation changes the selected profile; activation follows that profile's normal reload/restart policy.
|
|
131
131
|
|
|
@@ -145,7 +145,9 @@ If preferences cannot be read, available account rows remain visible but read-on
|
|
|
145
145
|
|
|
146
146
|
The managed GitHub Copilot group includes **Auto · Balance**, **Auto · Efficiency**, and **Auto · Intelligence** as flat virtual picker rows. They share one account-verified eligible pool after applying exact model exclusions and input-fit filtering. Routing now uses the account's **Powerful / Versatile / Lightweight** classification and conservative contextual task assessment, not reasoning effort or context capacity as a quality ranking. Even Intelligence uses Lightweight for an isolated greeting; short difficult reasoning stays Powerful. Unknown tasks prefer Versatile with Balance/Efficiency and Powerful with Intelligence. Suitable previous-model continuity wins before stable equal-weight category tie-breaking. Missing categories and fallback remain explicit. If no candidate fits, Auto preserves largest-input-budget native recovery; if none is eligible, it fails closed. One decision stays fixed across the turn's tool steps, retries and compaction. Images retain their independent verified admission. Search, actual model/Usage and histories remain native-owned. See the [task policy and limits](./docs/auto-task-routing.md) and [parent-inheritance boundary](./docs/automatic-model-routing.md#requested-parent-to-child-selection-inheritance). Core's picker remains flat.
|
|
147
147
|
|
|
148
|
-
|
|
148
|
+
Local rules also recognize isolated fenced JSON/CSV conversions and line sorting as routine, and exact continuations of a recent explicit difficult task as complex. New tasks break continuation inheritance. `uncertain` means demand is unresolved; insufficient local evidence is not proof that the conversation has no context.
|
|
149
|
+
|
|
150
|
+
The optional `github-copilot.autoSemanticAssessment: true` experiment recognizes contextual tasks with one bounded auxiliary model call **only when local demand is unknown**; it is **off by default** until calibrated. Its bounded input prioritizes current and preceding user requests without dismissing omitted context. It adds latency and supplier charges not included in native chat Usage. Failure or omitted assessment context never means the task is simple; it retains conservative local evidence. This is a local routing approximation, not GitHub's private Auto service.
|
|
149
151
|
|
|
150
152
|
Completed Copilot replies show **Auto (preference)** and an expandable recorded reason after native Usage/time, or **Manual** only with captured explicit fixed-selection evidence. A successful read with no retained evidence shows **Selection unknown**, never a guess from today's picker. Failed reads instead show **Selection unavailable → Retry**; retry only reads the same turn's record, not another model request. Explicit Session/turn arguments remain stable inside a session-bound Chat, avoiding the former native gateway argument-count failure on fresh master/lead replies. There is no repeated model name or separate Model details button: actual model display stays in native Usage, whose missing-route behavior is unchanged. New selection records are bounded and Host-local; restart, disposal or eviction loses them. Compatible historical Auto records remain readable. The public assistant-actions slot supports internal wrapping only, not changing the native fixed-height row.
|
|
151
153
|
|
|
@@ -184,7 +186,7 @@ Agents should treat the browser authorization as a human handoff, not as a token
|
|
|
184
186
|
5. Confirm **Signed in** and inspect the automatic discovery result before asking the user to choose a model. Already-signed-in Models opening ensures missing/stale metadata automatically; fresh ready cache makes no request. Use visible **Retry** for errors or **Manage → Refresh models** for an intentional forced update, not routine setup. Status alone does not discover, and login, metadata and successful model calls remain separate evidence.
|
|
185
187
|
6. Use **Sign out** only when the user explicitly asks to disconnect the account. It deletes the Copilot credential record but preserves route settings.
|
|
186
188
|
|
|
187
|
-
GitHub Releases and npm are the default distribution channels for each new version, using the same verified tarball. Pin the version and verify the evidence for the channel used: Release `SHA256SUMS`, and npm `dist.integrity` when installing from npm. Install into Desktop profiles only through Desktop's native package manager; the reserved `desktop` profile is not managed by the DSH CLI. After npm publication is verified, the native manager accepts `dsh-github-copilot@0.4.0-alpha.
|
|
189
|
+
GitHub Releases and npm are the default distribution channels for each new version, using the same verified tarball. Pin the version and verify the evidence for the channel used: Release `SHA256SUMS`, and npm `dist.integrity` when installing from npm. Install into Desktop profiles only through Desktop's native package manager; the reserved `desktop` profile is not managed by the DSH CLI. After npm publication is verified, the native manager accepts `dsh-github-copilot@0.4.0-alpha.83`, not a URL or file. Controlled offline CLI maintenance is documented for standalone named profiles only, never Desktop-managed profiles. Follow the mandatory search-composition preflight, backup, single-writer and post-install checks; do not bypass a corporate registry ban, disable TLS verification or restart without separate approval. See [distribution and publication requirements](./docs/npm-distribution.md).
|
|
188
190
|
|
|
189
191
|
**Desktop package-update and Web-lifecycle boundaries:** the native manager's `restart-required` result for an in-place installed-package upgrade is authoritative and distinct from plugin-bundle removal/re-add behavior. Any disable, removal, or upgrade involving service recomposition (`cordis.patch.yml`, such as the isolated routed web search layer) may leave dependent entries pending until Desktop is fully exited and cold-restarted. If the native manager reports only entries such as `pending (waiting for service: web)` while this plugin already shows **Off**, the requested disabled state has been persisted; cold-restart Desktop instead of repeatedly toggling or reinstalling the plugin. Other errors are not covered by this guidance and must still be investigated.
|
|
190
192
|
|
|
@@ -466,8 +468,8 @@ Report the published Release URL, version, tag/commit and verified asset SHA-256
|
|
|
466
468
|
`package.json` declares public npm distribution. A release tag must equal `v${package.json.version}`. Versions use standard SemVer prerelease labels (`alpha`, `beta`, or `rc`), each with its matching npm dist-tag; only stable versions use `latest`. The Release workflow performs the frozen install and complete verification gate, packs once (or recovers the original archive on retry), verifies `SHA256SUMS`, publishes the immutable GitHub Release and then publishes those same bytes to npm through OIDC. Either channel failing means delivery is incomplete. First package creation needs an authorized maintainer; staging requires an existing package and is not a first-package bootstrap. Historical releases are not republished.
|
|
467
469
|
|
|
468
470
|
```sh
|
|
469
|
-
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.
|
|
470
|
-
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.
|
|
471
|
+
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.83/dsh-github-copilot-0.4.0-alpha.83.tgz
|
|
472
|
+
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.83/SHA256SUMS
|
|
471
473
|
sha256sum --check SHA256SUMS
|
|
472
474
|
```
|
|
473
475
|
|
|
@@ -475,7 +477,7 @@ PowerShell can verify the same two downloaded files with:
|
|
|
475
477
|
|
|
476
478
|
```powershell
|
|
477
479
|
$expected = (Get-Content .\SHA256SUMS).Split()[0]
|
|
478
|
-
$actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.
|
|
480
|
+
$actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.83.tgz -Algorithm SHA256).Hash.ToLowerInvariant()
|
|
479
481
|
if ($actual -cne $expected) { throw 'Release checksum mismatch' }
|
|
480
482
|
```
|
|
481
483
|
|
package/README.zh.md
CHANGED
|
@@ -27,7 +27,7 @@ Core 管理。详见[图片兼容性与限制](./docs/image-input-compatibility.
|
|
|
27
27
|
|
|
28
28
|
截图来自当前构建 Client 的隔离浏览器 fixture,使用合成 Provider。它只证明已发布配置界面的形态,不证明真实凭据、账号权限、搜索成功或当前 Desktop 已加载。
|
|
29
29
|
|
|
30
|
-
> 当前文档与安装命令对应版本 `0.4.0-alpha.
|
|
30
|
+
> 当前文档与安装命令对应版本 `0.4.0-alpha.83`。源码、发布制品、已安装版本和实际加载运行时需分别确认;本地升级和中断会话的重启仍需用户批准。
|
|
31
31
|
|
|
32
32
|
Web search 路由现在优先显示在插件详情页;宿主不支持详情页插槽时才回退到 **设置 → 模型**或独立 Web search section。搜索子 Fiber 仍独立等待 Remote,不影响账号控件和既有搜索安全检查。
|
|
33
33
|
|
|
@@ -124,10 +124,10 @@ node package/scripts/check-search-composition.mjs --profile-dir /absolute/profil
|
|
|
124
124
|
仅对独立的具名 profile,获准且网络可用时,可通过受支持的 CLI 命令安装:
|
|
125
125
|
|
|
126
126
|
```sh
|
|
127
|
-
dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.
|
|
127
|
+
dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.83/dsh-github-copilot-0.4.0-alpha.83.tgz
|
|
128
128
|
```
|
|
129
129
|
|
|
130
|
-
Desktop profile 请在完成同样的预检后,使用 Desktop 原生包管理器安装 `dsh-github-copilot@0.4.0-alpha.
|
|
130
|
+
Desktop profile 请在完成同样的预检后,使用 Desktop 原生包管理器安装 `dsh-github-copilot@0.4.0-alpha.83`。Desktop 保留 `desktop` profile,CLI 不负责管理。若 registry 被公司封禁或不可用,不要更换网络绕行或使用离线 CLI;请停止。[受控离线 CLI 流程](./docs/npm-distribution.md#controlled-offline-cli-maintenance-for-standalone-profiles)仅适用于独立具名 profile,不适用于 Desktop。
|
|
131
131
|
|
|
132
132
|
随后打开上表对应的 Models UI,找到 **GitHub Copilot**,点击 **Sign in** 并完成 GitHub device-code 流程。安装会修改指定 profile;是否立即激活取决于该 profile 的常规 reload/restart 策略。
|
|
133
133
|
|
|
@@ -147,7 +147,9 @@ Desktop profile 请在完成同样的预检后,使用 Desktop 原生包管理
|
|
|
147
147
|
|
|
148
148
|
托管 GitHub Copilot 分组提供三个平铺的虚拟选择:**Auto · Balance**(均衡)、**Auto · Efficiency**(效率)和 **Auto · Intelligence**(智能)。三档共用精确排除与输入适配后的账号合格模型池。现在依据账号返回的 **Powerful / Versatile / Lightweight** 分类和保守的上下文任务判断选择,不再把推理档位或上下文容量当作品质排名。即使 Intelligence,独立问候也用 Lightweight;简短但困难的推理仍用 Powerful。无法判断的任务,Balance/Efficiency 优先 Versatile,Intelligence 优先 Powerful。上一轮模型仍符合目标分类与输入要求时优先保持它,否则仅在同类候选间稳定等权选择。分类缺失、分类回退均明确说明。没有可容纳候选时保留最大输入预算的原生恢复;没有合格候选则明确失败。同轮工具步骤、重试和自动压缩保持同一决策。图片验证独立保留;搜索、真实模型/Usage 和历史仍由原有服务所有。参见[任务策略与边界](./docs/auto-task-routing.md)及[父子继承边界](./docs/automatic-model-routing.md#requested-parent-to-child-selection-inheritance)。Core picker 仍保持平铺。
|
|
149
149
|
|
|
150
|
-
|
|
150
|
+
本地规则也会将独立、单一代码围栏内的 JSON/CSV 转换和行排序识别为常规任务,将近期明确困难任务的精确续接识别为复杂任务;新任务会切断继承。`uncertain` 表示需求尚无法确定,本地依据不足不代表会话没有上下文。
|
|
151
|
+
|
|
152
|
+
可选实验 `github-copilot.autoSemanticAssessment: true` **仅在本地需求未知时**通过一次有界的辅助模型请求判断上下文任务;校准前**默认关闭**。有界输入优先保留当前和前一条用户请求,但不会忽略上下文省略。它会增加延迟和供应方费用,不计入原生 Chat Usage。失败或判断上下文不完整不会被当作简单任务,而是保留保守本地依据。这是插件本地近似策略,不是 GitHub 的私有 Auto 服务。
|
|
151
153
|
|
|
152
154
|
已完成的 Copilot 回复在原生 Usage/时间后显示 **Auto(偏好)**,点信息按钮查看本轮记录的选择原因;只有捕获到明确固定选择才显示**手动**。成功读取但没有保留记录时显示**选择方式未知**,不借当前 picker 推断历史。读取失败则显示**选择记录读取失败 → 重试**;重试只重新读取同一轮记录,不重新调用模型。显式会话 ID/轮次参数不再被 Chat 的自动会话绑定缩减,修复新 master/lead 回复遭遇原生网关参数数量错误的问题。不重复模型名、不另设 Model details 按钮:实际模型仍由原生 Usage 显示,其缺失归属行为未改变。新选择记录仅限当前 Host 内存且有容量上限,重启、释放 Agent 或淘汰后会丢失;兼容历史 Auto 记录仍可读取。公开 assistant-actions slot 只允许插件内部换行,不能改变原生固定高度尾栏。
|
|
153
155
|
|
|
@@ -186,7 +188,7 @@ Agent 应把浏览器授权视为需要用户完成的 handoff,而不是自行
|
|
|
186
188
|
5. 确认 **Signed in** 并检查自动发现结果,再请用户选择模型。已登录时打开 Models 会自动确保缺失/过期元数据,新鲜 ready 缓存不发请求。错误可使用 **Retry**,有意强制更新时使用 **Manage → Refresh models**,不作为常规设置步骤。状态读取本身不发现;登录、元数据与真实调用成功是独立证据。
|
|
187
189
|
6. 只有用户明确要求断开账号时才使用 **Sign out**;它会删除 Copilot credential record,但保留 route settings。
|
|
188
190
|
|
|
189
|
-
每个新版本默认同时分发到 GitHub Releases 和 npm,两个渠道使用同一份已验证 tarball。应固定版本并核对所用渠道的证据:Release 的 `SHA256SUMS`;从 npm 安装时另核对 `dist.integrity`。Desktop profile 只能通过原生 Desktop 包管理器安装;获准且 registry 可用时输入 `dsh-github-copilot@0.4.0-alpha.
|
|
191
|
+
每个新版本默认同时分发到 GitHub Releases 和 npm,两个渠道使用同一份已验证 tarball。应固定版本并核对所用渠道的证据:Release 的 `SHA256SUMS`;从 npm 安装时另核对 `dist.integrity`。Desktop profile 只能通过原生 Desktop 包管理器安装;获准且 registry 可用时输入 `dsh-github-copilot@0.4.0-alpha.83`,不是 URL 或本地文件。Desktop 保留 `desktop` profile,不能通过 DSH CLI 管理。受控离线 CLI 维护仅适用于独立具名 profile,不得作为 Desktop 安装绕行方案。[双渠道发布与 OIDC 要求](./docs/npm-distribution.md)保持不变。
|
|
190
192
|
|
|
191
193
|
**Desktop package 更新与 Web 生命周期边界:**原生 manager 对已安装 package 原位升级返回的 `restart-required` 是权威结果,与移除/重加 plugin bundle 的行为是两个不同问题。涉及服务重编排补丁(`cordis.patch.yml`,如本插件所采用的隔离路由 Web 搜索层)的停用、移除或升级,可能让依赖条目保持 pending,直到完全退出并冷重启 Desktop。若原生 manager 只报告类似 `pending (waiting for service: web)` 的条目,同时本插件已经显示 **Off**,说明请求的停用状态已经落盘;请冷重启 Desktop,不要反复切换或重装插件。其它错误不属于此提示范围,仍需单独排查。
|
|
192
194
|
|
|
@@ -477,8 +479,8 @@ node scripts/agent.mjs attribution "DeepSeek Harness (DSH)"
|
|
|
477
479
|
`package.json` 声明公开 npm 分发。Release tag 必须严格等于 `v${package.json.version}`。预发布使用 `alpha`、`beta` 或 `rc` 及对应 npm dist-tag,只有稳定版使用 `latest`。Release workflow 执行 frozen install 和完整门禁,只打包一次(重试恢复原始归档),验证 `SHA256SUMS`,发布不可变 GitHub Release,再通过 OIDC 将同一份字节发布到 npm。任一渠道失败都表示交付未完成。首次建包须由获准环境中的维护者完成;staging 要求包已存在,不能代替首次建包。不会批量补发历史版本。
|
|
478
480
|
|
|
479
481
|
```sh
|
|
480
|
-
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.
|
|
481
|
-
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.
|
|
482
|
+
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.83/dsh-github-copilot-0.4.0-alpha.83.tgz
|
|
483
|
+
curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.83/SHA256SUMS
|
|
482
484
|
sha256sum --check SHA256SUMS
|
|
483
485
|
```
|
|
484
486
|
|
|
@@ -486,7 +488,7 @@ PowerShell 可以对已下载的同一组文件执行:
|
|
|
486
488
|
|
|
487
489
|
```powershell
|
|
488
490
|
$expected = (Get-Content .\SHA256SUMS).Split()[0]
|
|
489
|
-
$actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.
|
|
491
|
+
$actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.83.tgz -Algorithm SHA256).Hash.ToLowerInvariant()
|
|
490
492
|
if ($actual -cne $expected) { throw 'Release checksum mismatch' }
|
|
491
493
|
```
|
|
492
494
|
|
package/deployment-baseline.json
CHANGED
|
@@ -8,7 +8,7 @@
|
|
|
8
8
|
},
|
|
9
9
|
"package": {
|
|
10
10
|
"name": "dsh-github-copilot",
|
|
11
|
-
"version": "0.4.0-alpha.
|
|
11
|
+
"version": "0.4.0-alpha.83"
|
|
12
12
|
},
|
|
13
13
|
"evidence": {
|
|
14
14
|
"kind": "source-and-synthetic-test-inventory",
|
|
@@ -1111,6 +1111,10 @@
|
|
|
1111
1111
|
{"file": "tests/auto-model-routing.spec.ts", "name": "preserves suitable continuity but never lets it override task category or input fit"},
|
|
1112
1112
|
{"file": "tests/auto-model-routing.spec.ts", "name": "discloses category fallbacks and never guesses missing category from ID or capacity"},
|
|
1113
1113
|
{"file": "tests/auto-task-assessment.spec.ts", "name": "bounds context, never reads attachment/reasoning bodies, and does not downshift omitted context"},
|
|
1114
|
+
{"file": "tests/auto-task-assessment.spec.ts", "name": "inherits difficult-task evidence only through explicit bounded continuations"},
|
|
1115
|
+
{"file": "tests/auto-task-assessment.spec.ts", "name": "recognizes only isolated explicit fenced mechanical transformations as routine"},
|
|
1116
|
+
{"file": "tests/auto-task-assessment.spec.ts", "name": "preserves valid bounded rows and prioritizes the current user and nearest task over verbose output"},
|
|
1117
|
+
{"file": "tests/auto-selection-explanation.spec.ts", "name": "distinguishes local rule coverage from absent conversation context in both languages"},
|
|
1114
1118
|
{"file": "tests/auto-task-assessment.spec.ts", "name": "preserves caller cancellation instead of using a success-shaped fallback"},
|
|
1115
1119
|
{"file": "tests/preview-route.spec.ts", "name": "assesses Auto through one concrete public native adapter call and freezes the captured reason across retries"},
|
|
1116
1120
|
{"file": "tests/turn-selection-card.spec.ts", "name": "explains a captured task/category decision with progressive details (%s)"},
|
|
@@ -1,6 +1,7 @@
|
|
|
1
1
|
# Contextual Auto routing and captured selection reasons
|
|
2
2
|
|
|
3
|
-
Tracking: [#258](https://github.com/cloga/dsh-github-copilot/issues/258)
|
|
3
|
+
Tracking: [#258](https://github.com/cloga/dsh-github-copilot/issues/258) and
|
|
4
|
+
[evidence refinement #262](https://github.com/cloga/dsh-github-copilot/issues/262).
|
|
4
5
|
|
|
5
6
|
## Requirements
|
|
6
7
|
|
|
@@ -35,6 +36,20 @@ outcome, never silently translated into simple. Optional semantic assessment is
|
|
|
35
36
|
an experiment until evaluated against annotated tasks; it must not claim
|
|
36
37
|
equivalence to GitHub's private router or calibrated confidence.
|
|
37
38
|
|
|
39
|
+
Local rules recognize an isolated, explicit JSON-to-CSV/CSV-to-JSON conversion
|
|
40
|
+
or line sort only when the entire request is at most 1600 characters, contains
|
|
41
|
+
one fenced payload and no other work or non-text content. This establishes
|
|
42
|
+
routine, not simple. An exact acknowledgement/continue request can inherit
|
|
43
|
+
explicit difficult-task evidence from the nearest non-continuation user request
|
|
44
|
+
among the preceding twelve user requests. A new task or an expired anchor stops
|
|
45
|
+
inheritance; simple/routine demand is never inherited.
|
|
46
|
+
|
|
47
|
+
`uncertain` remains a valid task result. `insufficient-evidence` means the local
|
|
48
|
+
rules cannot establish demand, not that the conversation has no context.
|
|
49
|
+
Attachments, unfamiliar instructions and genuinely ambiguous continuations can
|
|
50
|
+
still remain unknown. These refinements do not change category preference,
|
|
51
|
+
previous-model eligibility or the default-disabled semantic experiment.
|
|
52
|
+
|
|
38
53
|
The semantic experiment uses one fixed concrete account candidate through the
|
|
39
54
|
existing managed native adapter, never an Auto ID and never `ctx.llm` routing
|
|
40
55
|
recursion. It has no tools, bounded text-only assessment context, bounded JSON
|
|
@@ -44,6 +59,14 @@ are not read; omitted context makes classification uncertain and cannot justify
|
|
|
44
59
|
downshifting. Assessment input is a separate projection: the actual chat
|
|
45
60
|
messages, system prompt, tools, attachments and history stay untouched.
|
|
46
61
|
|
|
62
|
+
Only locally unknown tasks invoke the enabled experiment. Context packing keeps
|
|
63
|
+
the current and nearest preceding user requests before recent output, then
|
|
64
|
+
restores chronological order. At most twelve complete JSON rows, 1600 text
|
|
65
|
+
characters per row and 8000 serialized characters are retained; JSON escape
|
|
66
|
+
expansion is included in this bound. Any truncation, skipped row or non-text
|
|
67
|
+
content still marks omitted context and prevents a semantic simple/routine
|
|
68
|
+
downshift. Prioritization does not prove omitted output irrelevant.
|
|
69
|
+
|
|
47
70
|
The strict result contains demand (`simple`, `routine`, `complex`, `unknown`)
|
|
48
71
|
and finite evidence signals, not a model ID, raw reasoning or an uncalibrated
|
|
49
72
|
confidence percentage. Invalid/truncated/failed results retain the local
|
|
@@ -52,11 +52,16 @@ Before applying soft preferences, routing checks whether each candidate can acco
|
|
|
52
52
|
### Task-aware supplier category preferences
|
|
53
53
|
|
|
54
54
|
All three preferences use the *same* hard-eligible account pool. The default local
|
|
55
|
-
assessment recognizes isolated greetings
|
|
56
|
-
investigation
|
|
55
|
+
assessment recognizes isolated greetings, explicit difficult reasoning or
|
|
56
|
+
investigation, bounded explicit continuations of those difficult tasks, and
|
|
57
|
+
isolated fenced mechanical transformations. Other tasks remain uncertain rather
|
|
58
|
+
than guessing from length; insufficient local evidence does not mean the
|
|
59
|
+
conversation lacks context.
|
|
57
60
|
An optional `github-copilot.autoSemanticAssessment: true` experiment makes one
|
|
58
61
|
bounded auxiliary native-adapter call to a concrete, non-excluded supplier
|
|
59
|
-
Lightweight model
|
|
62
|
+
Lightweight model only for locally unknown demand. It prioritizes the current
|
|
63
|
+
and preceding user requests without removing omitted-context safeguards. It
|
|
64
|
+
adds latency and supplier charges not included in native
|
|
60
65
|
chat Usage. It is off by default pending labeled-corpus calibration.
|
|
61
66
|
|
|
62
67
|
| Task demand | Efficiency | Balance (`auto`) | Intelligence |
|
package/lib/client.js
CHANGED
|
@@ -6330,12 +6330,12 @@ Set the \`cycles\` parameter to \`"ref"\` to resolve cyclical schemas with defs.
|
|
|
6330
6330
|
investigation: zh ? "调查或调试" : "investigation or debugging",
|
|
6331
6331
|
continuation: zh ? "继续已有工作" : "continuation of existing work",
|
|
6332
6332
|
"context-omitted": zh ? "判断上下文不完整" : "incomplete assessment context",
|
|
6333
|
-
"insufficient-evidence": zh ? "
|
|
6333
|
+
"insufficient-evidence": zh ? "本地规则尚不能确定任务需求" : "local rules cannot establish task demand"
|
|
6334
6334
|
};
|
|
6335
6335
|
return {
|
|
6336
6336
|
conclusion: explanation.method === "no-fit" ? zh ? "没有候选能直接容纳估算输入;选中输入预算最大的模型,保留原生压缩或失败处理。" : "No candidate fits the estimated input. Selected the largest input budget for native recovery or failure." : explanation.fallback ? zh ? `目标分类 ${category[explanation.targetCategory]} 没有合格的可容纳候选,回退到 ${category[explanation.selectedCategory]}。` : `No eligible fitting ${category[explanation.targetCategory]} candidate; fell back to ${category[explanation.selectedCategory]}.` : zh ? `本轮判断为${demand[explanation.assessment.demand]};按当前偏好选择 ${category[explanation.selectedCategory]} 类模型。` : `This task was assessed as ${demand[explanation.assessment.demand]}; the preference policy selected ${category[explanation.selectedCategory]}.`,
|
|
6337
6337
|
choice: explanation.method === "continuity" ? zh ? "上一轮模型仍满足本轮分类和输入要求,因此保持它。" : "Kept the previous model because it still meets this turn’s category and input requirements." : explanation.method === "only-candidate" ? zh ? "该分类只有一个可容纳的合格候选。" : "Only one eligible fitting candidate remains in this category." : explanation.method === "equal-distribution" ? zh ? `该分类有 ${explanation.categoryCandidateCount} 个可容纳候选;按本轮稳定等权规则选中,不代表质量排名。` : `Selected among ${explanation.categoryCandidateCount} fitting category candidates by stable equal-weight allocation, not a quality ranking.` : zh ? "所有候选均不适配;这不是成功的偏好匹配。" : "All candidates are oversized; this is not a successful preference match.",
|
|
6338
|
-
diagnostic: explanation.assessment.diagnostic === void 0 ? void 0 : explanation.assessment.diagnostic === "disabled" ? zh ? "
|
|
6338
|
+
diagnostic: explanation.assessment.diagnostic === void 0 ? void 0 : explanation.assessment.diagnostic === "disabled" ? zh ? "语义判断实验未开启;这是本地规则的判断边界,不代表会话没有上下文。" : "Semantic assessment is disabled; this is a limit of local rules, not proof that the conversation lacks context." : zh ? `语义判断未提供完整可用依据(${{
|
|
6339
6339
|
unavailable: "没有可用分类模型",
|
|
6340
6340
|
"invalid-result": "结果无效",
|
|
6341
6341
|
timeout: "超时",
|