dsh-github-copilot 0.4.0-alpha.97 → 0.4.0-alpha.99

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/AGENTS.md CHANGED
@@ -68,7 +68,7 @@ Alpha.25 admits native `subagent/descriptor` v3, already v3 in the retained rc.1
68
68
  - `src/responses-reasoning.ts`, `src/responses-reasoning-text.ts`: selected-model effort mapping and public summary assembly.
69
69
  - `src/tool-schema-compat.ts`: Copilot-only prompt-assembly filter for unusable escalation arguments and action-specific Goal update schemas.
70
70
  - `src/plan.ts`: Copilot-only, fail-closed hosted-search candidate lifecycle.
71
- - `src/auto-assessment-evidence.ts`: bounded request-local semantic classifier milestones copied into strict ephemeral turn evidence; no prompts, replay, credentials or error bodies. Total includes preparation; adapter start and first text are not HTTP dispatch/byte timings. Keep one 8s deadline and 128-token compact output. Request reasoning-off only when supplier and public prepared-model effort agree; unsupported native controls stay unchanged. No guessed latency ranking, adaptive timeout, cooldown or cross-turn cache without reviewed observations.
71
+ - `src/auto-assessment-evidence.ts`: bounded request-local semantic classifier milestones copied into strict ephemeral turn evidence; no prompts, replay, credentials or error bodies. Total includes preparation; adapter start and first text are not HTTP dispatch/byte timings. Keep one 8s monotonic deadline checked at preparation/native request/result boundaries and 128-token compact output; delayed timer settlement is not a real-time guarantee. Prefer eligible supplier-advertised reasoning-off classifiers, but request off only when supplier and public prepared-model effort agree; unsupported native controls stay unchanged. No guessed latency ranking, adaptive timeout, cooldown or cross-turn cache without reviewed observations.
72
72
  - `src/probe.ts`: bounded native-search capability proof.
73
73
  - `src/wire.ts`, `src/wire-anthropic.ts`: inline hosted-search streaming.
74
74
  - `src/traditional-search.ts`: `github-copilot-hosted` `ctx.web` provider.
@@ -285,6 +285,8 @@ Capture the traced `remote.githubCopilot` namespace once per account UI registra
285
285
 
286
286
  Auto history compatibility (alpha.54, #204): official rc.2 `Session.append()` has no public ignorable-envelope option. Do not emit optional `github-copilot/auto-model-decision` events or borrow another event type; retain native real-model provenance and durable Auto selection without new durable decision recording. The assistant-actions selection footer reads bounded Host-lifetime decisions through native agent-scoped lookup and compatible historical events. Explicit captured fixed selection alone proves Manual; missing evidence is unknown, never reconstructed from today's picker. Do not repeat model names, add Model details, change native Usage/context metering or modify parent DOM. `scripts/repair-auto-model-history.mjs` is explicit source-checkout maintenance, not a runtime hook: strict pinned official v4 validation, check-only by default, optional detached repaired copy plus byte-exact backup. It never replaces live history. Applying a copy requires separately approved stopped writers and a fresh source-hash check. See `docs/automatic-model-routing.md`.
287
287
 
288
+ Auto intent continuity (#293) uses the strict `githubCopilotAutoModelIntent` public Session projection of existing owner-explicit `model/selection` events. Execution headers cannot erase intent; fixed/other-provider choices replace it for future turns. Ignore copied fork/child prefixes, retain native pending and parent-follow precedence, and diagnose missing/invalid evidence. Freeze admitted Auto route configuration across late choices without writing history, settings or historical reasons. Projection disposal must remove only its own registration.
289
+
288
290
  1. Identify the owning seam using `agent-contract.json`; do not duplicate an upstream owner.
289
291
  2. Add or update focused tests before changing deployment claims.
290
292
  3. Update both READMEs when user behavior, setup, migration, or boundaries change.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,17 @@
1
1
  # Changelog
2
2
 
3
+ ## 0.4.0-alpha.99 (prepared)
4
+
5
+ - Enforce Auto assessment's single monotonic deadline at preparation, native request and result boundaries even when a delayed timer has not fired (#295). Dispose the assessment timer and listener on settlement; preserve cancellation, proof revocation and unknown-demand fallback.
6
+ - Prefer eligible supplier Lightweight classifiers advertising reasoning-off, with deterministic ID tie-breaking. Send off only with matching published native support, without latency guesses, another classifier or retries.
7
+ - Explain the fitting category pool and continuity separately from semantic model merit. Preserve previous-model policy, turn freezing, native Usage and history; no cross-turn cache or Core changes.
8
+
9
+ ## 0.4.0-alpha.98 (prepared)
10
+
11
+ - Retain owner-explicit Auto intent through a strict public Session projection when a native virtual request header consumes pending selection (#293).
12
+ - Preserve intent across concrete execution headers and cold replay; later fixed/other-provider choices replace it for future turns. Admitted Auto requests keep their route through late picker changes.
13
+ - Ignore copied fork/child prefixes as owner-explicit evidence; keep native pending/parent-follow precedence and named missing/invalid diagnostics. No Core changes, new durable events, settings/history rewrite or fabricated historical reasons.
14
+
3
15
  ## 0.4.0-alpha.97 (prepared)
4
16
 
5
17
  - Add request-scoped public native body-write/header timing and optional TLS ALPN/Node buffer observations to exact verified Copilot request-body timeout diagnostics (#291).
package/README.md CHANGED
@@ -8,7 +8,7 @@
8
8
 
9
9
  GitHub Copilot account discovery, contextual Auto routing and hosted search for DSH. The plugin reuses DSH's published native adapter and one Host-owned OAuth account; it does not patch Core or maintain a second model catalog.
10
10
 
11
- **Source candidate: `0.4.0-alpha.97` (unreleased). Supported host: official DSH / Windows Desktop `0.2.0-rc.2`.** Earlier DSH pins are historical evidence, not supported installation targets. Publication, profile installation and the version loaded by a running Host are separate states. The versioned commands below are for this candidate after publication, not evidence that its assets exist.
11
+ **Source candidate: `0.4.0-alpha.99` (unreleased). Supported host: official DSH / Windows Desktop `0.2.0-rc.2`.** Earlier DSH pins are historical evidence, not supported installation targets. Publication, profile installation and the version loaded by a running Host are separate states. The versioned commands below are for this candidate after publication, not evidence that its assets exist.
12
12
 
13
13
  ## What you can do
14
14
 
@@ -40,10 +40,10 @@ Supply any launcher patches with repeated `--patch /absolute/file`. Require `sup
40
40
  For a **standalone named profile**:
41
41
 
42
42
  ```sh
43
- dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.97/dsh-github-copilot-0.4.0-alpha.97.tgz
43
+ dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.99/dsh-github-copilot-0.4.0-alpha.99.tgz
44
44
  ```
45
45
 
46
- For **Desktop**, its native package manager accepts `dsh-github-copilot@0.4.0-alpha.97` after publication. Official rc.2's **Desktop-bundled CLI** also supports reserved-profile plugin management; a global/generic `dsh` shim is not equivalent. Qualify the installed entry before use. See [Desktop CLI qualification](./docs/npm-distribution.md#desktop-bundled-cli-on-official-rc2) and the distinct [standalone offline procedure](./docs/npm-distribution.md#controlled-offline-cli-maintenance-for-standalone-profiles). Neither path authorizes registry-policy bypass, peer patches or configuration deletion.
46
+ For **Desktop**, its native package manager accepts `dsh-github-copilot@0.4.0-alpha.99` after publication. Official rc.2's **Desktop-bundled CLI** also supports reserved-profile plugin management; a global/generic `dsh` shim is not equivalent. Qualify the installed entry before use. See [Desktop CLI qualification](./docs/npm-distribution.md#desktop-bundled-cli-on-official-rc2) and the distinct [standalone offline procedure](./docs/npm-distribution.md#controlled-offline-cli-maintenance-for-standalone-profiles). Neither path authorizes registry-policy bypass, peer patches or configuration deletion.
47
47
 
48
48
  After an approved reload/restart:
49
49
 
@@ -70,12 +70,16 @@ Categories come from authenticated supplier metadata, not model names, context c
70
70
 
71
71
  Semantic assessment is **enabled by default only for locally unknown tasks**. It makes at most one bounded auxiliary call with an **8-second end-to-end deadline** and **128-token output budget**. Set `github-copilot.autoSemanticAssessment: false` to opt out. Timeout/invalid output leaves demand unknown and uses the preference fallback; it never means simple or triggers another classifier. Caller cancellation and account invalidation remain terminal. Auxiliary calls add latency and supplier charges outside native answer Usage.
72
72
 
73
+ The deadline is checked against a monotonic clock at preparation, native request and result boundaries, even if the event loop delays the timer. Such delays can postpone settlement; they do not authorize late dispatch or late-result acceptance. Eligible Lightweight classifiers advertising reasoning `off` take precedence, with deterministic ID tie-breaking; the request uses `off` only if the public native model also supports it. This is not a measured speed ranking. Explanations distinguish the fitting category pool and previous-model continuity from semantic merit.
74
+
73
75
  The reply's selection disclosure shows captured reasons and optional content-free auxiliary milestones. `uncertain` and `insufficient-evidence` do not mean the conversation is empty. **Selection unknown** means no retained record; **Selection unavailable → Retry** means the read failed. Retry rereads that turn, not inference. Evidence is bounded and Host-local, so restart/eviction can lose it.
74
76
 
75
77
  **Model exclusions** remove exact IDs from the managed picker and new Auto pools. Selected models are excludable: an admitted turn keeps its model, but a new turn/direct request cannot use an excluded ID. A fixed selection is never silently replaced. Unknown preference settings leave controls read-only, not falsely enabled. Saves use narrow native CAS without credentials, discovery or Session scans.
76
78
 
77
79
  See [task assessment](./docs/auto-task-routing.md) and [routing contracts, exclusions and history recovery](./docs/automatic-model-routing.md).
78
80
 
81
+ Explicit Auto intent survives native pending-selection consumption and cold restore through a plugin-owned projection of existing selection events. Later explicit fixed choices still win for future turns; admitted turns keep their route. This does not backfill historical selection reasons or turn unknown evidence into guessed Auto.
82
+
79
83
  ## Subagents and search
80
84
 
81
85
  ![Current parent-following and Web search controls](./docs/images/copilot-search-routing.png)
@@ -177,14 +181,14 @@ Commit attribution uses `Assisted-by` with the actual tool, never the model prov
177
181
  GitHub Releases and npm distribute the same original verified tarball. Pin a version; verify Release SHA-256 or npm `dist.integrity`. Never repack an immutable release or move/reuse its tag.
178
182
 
179
183
  ```sh
180
- curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.97/dsh-github-copilot-0.4.0-alpha.97.tgz
181
- curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.97/SHA256SUMS
184
+ curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.99/dsh-github-copilot-0.4.0-alpha.99.tgz
185
+ curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.99/SHA256SUMS
182
186
  sha256sum --check SHA256SUMS
183
187
  ```
184
188
 
185
189
  ```powershell
186
190
  $expected = (Get-Content .\SHA256SUMS).Split()[0]
187
- $actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.97.tgz -Algorithm SHA256).Hash.ToLowerInvariant()
191
+ $actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.99.tgz -Algorithm SHA256).Hash.ToLowerInvariant()
188
192
  if ($actual -cne $expected) { throw 'Release checksum mismatch' }
189
193
  ```
190
194
 
package/README.zh.md CHANGED
@@ -8,7 +8,7 @@
8
8
 
9
9
  为 DSH 提供 GitHub Copilot 账号模型发现、上下文 Auto 选模和 hosted search。插件复用 DSH 公开的原生适配器与一个 Host 端 OAuth 账号,不修改 Core,也不维护第二套模型目录。
10
10
 
11
- **源码候选版本:`0.4.0-alpha.97`(尚未发布)。支持宿主:官方 DSH / Windows Desktop `0.2.0-rc.2`。** 较早 DSH pin 仅是历史证据,不是当前安装目标。已发布、已安装到 profile、已被运行中的 Host 加载,是三个不同状态。下方带版本的命令适用于候选版本发布后,不代表对应资产已经存在。
11
+ **源码候选版本:`0.4.0-alpha.99`(尚未发布)。支持宿主:官方 DSH / Windows Desktop `0.2.0-rc.2`。** 较早 DSH pin 仅是历史证据,不是当前安装目标。已发布、已安装到 profile、已被运行中的 Host 加载,是三个不同状态。下方带版本的命令适用于候选版本发布后,不代表对应资产已经存在。
12
12
 
13
13
  ## 你可以做什么
14
14
 
@@ -40,10 +40,10 @@ node package/scripts/check-search-composition.mjs --profile-dir /absolute/profil
40
40
  **独立具名 profile**:
41
41
 
42
42
  ```sh
43
- dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.97/dsh-github-copilot-0.4.0-alpha.97.tgz
43
+ dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.99/dsh-github-copilot-0.4.0-alpha.99.tgz
44
44
  ```
45
45
 
46
- **Desktop** 原生包管理器在发布后接受 `dsh-github-copilot@0.4.0-alpha.97`。官方 rc.2 **Desktop 随附的专属 CLI**也支持管理保留 profile;全局/普通 `dsh` shim 不等价。使用前核验实际安装入口,参见[Desktop CLI 核验](./docs/npm-distribution.md#desktop-bundled-cli-on-official-rc2)及独立的[具名 profile 离线流程](./docs/npm-distribution.md#controlled-offline-cli-maintenance-for-standalone-profiles)。两种路径都不授权绕过包源政策、修改 peer 或删除配置。
46
+ **Desktop** 原生包管理器在发布后接受 `dsh-github-copilot@0.4.0-alpha.99`。官方 rc.2 **Desktop 随附的专属 CLI**也支持管理保留 profile;全局/普通 `dsh` shim 不等价。使用前核验实际安装入口,参见[Desktop CLI 核验](./docs/npm-distribution.md#desktop-bundled-cli-on-official-rc2)及独立的[具名 profile 离线流程](./docs/npm-distribution.md#controlled-offline-cli-maintenance-for-standalone-profiles)。两种路径都不授权绕过包源政策、修改 peer 或删除配置。
47
47
 
48
48
  经批准 reload/restart 后:
49
49
 
@@ -70,12 +70,16 @@ dsh plugin --profile web add https://github.com/cloga/dsh-github-copilot/release
70
70
 
71
71
  语义判断**默认开启,仅用于本地需求未知的任务**。最多一次辅助请求,**8 秒端到端期限、128 tokens 输出预算**。可用 `github-copilot.autoSemanticAssessment: false` 关闭。超时/无效结果保留未知,再按偏好兜底,不解释为简单,也不换分类模型重试。用户取消和账号失效仍终止请求。辅助调用增加延迟与供应方费用,不计入原生回答 Usage。
72
72
 
73
+ 准备、原生请求和结果边界均用单调时钟检查期限,不依赖定时器及时执行。事件循环延迟可能推迟结束,但不授权超期发起请求或接受结果。合格 Lightweight 分类模型优先考虑声明 reasoning `off` 的候选,同级按 ID 确定顺序;只有公开原生模型也支持时才发送 `off`。这不是实测速率排名。解释分别说明该分类的可容纳候选数和上一模型连续性,不将保留上一模型说成语义优选。
74
+
73
75
  回复的选模说明展示捕获的理由和可选的不含内容的辅助里程碑。`uncertain`、`insufficient-evidence` 不等于会话没有上下文。**Selection unknown**表示记录未保留,**Selection unavailable → Retry**表示读取失败;Retry 只重读该 turn,不重新推理。证据有界且仅限 Host 生命周期,重启/淘汰可能丢失。
74
76
 
75
77
  **模型排除**将精确 ID 移出托管 picker 与新 turn 的 Auto 候选。已选模型也能排除:已准入 turn 保持原模型,但新 turn/直接调用不得使用;固定选择不会静默替换。偏好状态未知时只读,不伪装为已启用。保存采用窄范围原生 CAS,不读凭据、不发现模型、不扫描会话。
76
78
 
77
79
  详见[任务判断](./docs/auto-task-routing.md)与[路由合同、排除和历史恢复](./docs/automatic-model-routing.md)。
78
80
 
81
+ 显式 Auto 意图通过插件对现有选择事件的独立投影保留,不因原生 pending 被消费或冷恢复而丢失。后续显式固定选择仍从未来轮次生效,已准入轮次保持原路由。不会补写历史选择理由,也不会把未知证据猜成 Auto。
82
+
79
83
  ## 子代理与搜索
80
84
 
81
85
  ![当前父模型跟随及 Web search 控件](./docs/images/copilot-search-routing.png)
@@ -177,14 +181,14 @@ Agent 从 `node scripts/agent.mjs describe --json`、`doctor --json`及`plan <ta
177
181
  GitHub Releases 和 npm 分发同一原始已校验 tarball。固定版本并核验 Release SHA-256 或 npm `dist.integrity`;不重打包不可变 Release、不移动/复用 tag。
178
182
 
179
183
  ```sh
180
- curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.97/dsh-github-copilot-0.4.0-alpha.97.tgz
181
- curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.97/SHA256SUMS
184
+ curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.99/dsh-github-copilot-0.4.0-alpha.99.tgz
185
+ curl -LO https://github.com/cloga/dsh-github-copilot/releases/download/v0.4.0-alpha.99/SHA256SUMS
182
186
  sha256sum --check SHA256SUMS
183
187
  ```
184
188
 
185
189
  ```powershell
186
190
  $expected = (Get-Content .\SHA256SUMS).Split()[0]
187
- $actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.97.tgz -Algorithm SHA256).Hash.ToLowerInvariant()
191
+ $actual = (Get-FileHash .\dsh-github-copilot-0.4.0-alpha.99.tgz -Algorithm SHA256).Hash.ToLowerInvariant()
188
192
  if ($actual -cne $expected) { throw 'Release checksum mismatch' }
189
193
  ```
190
194
 
@@ -8,7 +8,7 @@
8
8
  },
9
9
  "package": {
10
10
  "name": "dsh-github-copilot",
11
- "version": "0.4.0-alpha.97"
11
+ "version": "0.4.0-alpha.99"
12
12
  },
13
13
  "evidence": {
14
14
  "kind": "source-and-synthetic-test-inventory",
@@ -584,6 +584,7 @@
584
584
  {"file": "src/remote.ts", "marker": "githubCopilot.setModelExcluded"},
585
585
  {"file": "src/preview-provider.ts", "marker": "native.streamSimple"},
586
586
  {"file": "src/auto-assessment-evidence.ts", "marker": "SemanticAssessmentEvidenceSchema"},
587
+ {"file": "src/auto-task-assessment.ts", "marker": "performance.now() >= deadline"},
587
588
  {"file": "src/request-body-timeout-marker.ts", "marker": "COPILOT_REQUEST_BODY_TIMEOUT"},
588
589
  {"file": "src/request-body-evidence.ts", "marker": "REQUEST_BODY_EVIDENCE_LIMITS"},
589
590
  {"file": "src/request-upload-evidence.ts", "marker": "createRequestUploadObserver"},
@@ -593,6 +594,12 @@
593
594
  {"file": "src/pi-provider-bridge.ts", "marker": "PiStreamPublicCompatibility"}
594
595
  ],
595
596
  "tests": [
597
+ {"file": "tests/auto-task-assessment.spec.ts", "name": "rejects late results even before the scheduled timeout callback runs"},
598
+ {"file": "tests/auto-task-assessment.spec.ts", "name": "aborts preparation at the monotonic deadline before an adapter can start"},
599
+ {"file": "tests/preview-route.spec.ts", "name": "does not dispatch an auxiliary request when native preparation outlasts a delayed deadline timer"},
600
+ {"file": "tests/preview-provider.spec.ts", "name": "checks the auxiliary deadline immediately before native %s Fetch and releases the lease"},
601
+ {"file": "tests/auto-task-classifier.spec.ts", "name": "prefers advertised reasoning-off candidates without inferring latency or model-name capabilities"},
602
+ {"file": "tests/auto-selection-explanation.spec.ts", "name": "distinguishes continuity and category pool from semantic model merit (%s)"},
596
603
  {"file": "tests/preview-provider.spec.ts", "name": "routes an unseen ID through published Core and native %s with Bearer auth"},
597
604
  {"file": "tests/preview-provider.spec.ts", "name": "takes protocol and independent input limits from descriptor evidence rather than the local GPT-6 catalog"},
598
605
  {"file": "tests/preview-provider.spec.ts", "name": "does not infer controls from a name or unfamiliar advertised effort labels"},
@@ -1122,6 +1129,23 @@
1122
1129
  {"file": "tests/client-no-model-roles.spec.ts", "name": "does not resurrect a role entry when a retained legacy Remote arrives later"}
1123
1130
  ]
1124
1131
  },
1132
+ {
1133
+ "id": "managed-copilot-auto-intent-continuity",
1134
+ "required": true,
1135
+ "sourceMarkers": [
1136
+ {"file": "src/auto-model-intent.ts", "marker": "githubCopilotAutoModelIntent"},
1137
+ {"file": "src/auto-model-intent.ts", "marker": "state.inherited"},
1138
+ {"file": "src/auto-model-host.ts", "marker": "COPILOT_AUTO_INTENT_UNAVAILABLE"},
1139
+ {"file": "src/auto-model-host.ts", "marker": "removeIntentProjection"},
1140
+ {"file": "docs/automatic-model-routing.md", "marker": "Cold refolding restores intent"}
1141
+ ],
1142
+ "tests": [
1143
+ {"file": "tests/auto-model-intent.spec.ts", "name": "retains explicit intent across virtual and concrete execution headers and cold replay"},
1144
+ {"file": "tests/auto-model-intent.spec.ts", "name": "does not infer intent from execution or copied fork/child prefixes"},
1145
+ {"file": "tests/auto-model-host.spec.ts", "name": "diagnoses unavailable continuity evidence ($error) without blocking a new explicit fixed choice"},
1146
+ {"file": "tests/fixtures/session-context-core.fixture.ts", "name": "retains explicit Auto after native virtual-header consumption without relabeling prior turns"}
1147
+ ]
1148
+ },
1125
1149
  {
1126
1150
  "id": "managed-copilot-auto-model-routing",
1127
1151
  "required": true,
@@ -3,7 +3,7 @@
3
3
  Tracking: [#258](https://github.com/cloga/dsh-github-copilot/issues/258) and
4
4
  [evidence refinement #262](https://github.com/cloga/dsh-github-copilot/issues/262).
5
5
 
6
- **Current contract:** published `0.4.0-alpha.88`, targeting official DSH/Windows
6
+ **Current contract:** source candidate `0.4.0-alpha.99`, targeting official DSH/Windows
7
7
  Desktop `0.2.0-rc.2`. Semantic assessment is enabled by default with bounded
8
8
  auxiliary timing/output evidence. Neither publication nor
9
9
  these explanations prove the version loaded in a particular Host.
@@ -123,6 +123,18 @@ retains its seven-signal compatibility bound; raw explanation, confidence
123
123
  percentages and reasoning are not requested. Truncation remains failure, never
124
124
  accepted partial JSON.
125
125
 
126
+ One monotonic deadline is checked after context packing, discovery and native
127
+ preparation, before native request admission, on delivered chunks and before
128
+ accepting a decoded result. A delayed event-loop timer cannot authorize a
129
+ late auxiliary request or result. It can still delay observed settlement:
130
+ this is not a hard real-time guarantee or evidence of a supplier latency cause.
131
+ The owned timer and abort listener are disposed when assessment settles.
132
+
133
+ Among eligible supplier Lightweight text candidates with the existing input
134
+ headroom, prefer advertised reasoning `off`, then deterministic ID ordering.
135
+ This does not infer speed, health or quality from effort or model names, and
136
+ does not launch a second candidate if native support is absent.
137
+
126
138
  Request `off` reasoning only when both current supplier metadata advertises it
127
139
  and the public prepared model lists an `off` effort. Supplier metadata alone
128
140
  does not establish native support: current managed reasoning maps can decline
@@ -147,6 +159,8 @@ and frozen in the bounded store and routing decision, not durably appended or
147
159
  reconstructed after restart. The footer says assessment timed out, demand
148
160
  remained unknown, and preference policy selected the category as a fallback.
149
161
  Progressive details show milestones separately from native answer Usage.
162
+ Continuity copy states the fitting category-pool count and explicitly disclaims
163
+ semantic ranking; all-input-fit counts are not the final category pool.
150
164
 
151
165
  This is instrumentation and bounded output optimization, not a measured latency
152
166
  improvement or classifier calibration. No latency-ranked classifier pool,
@@ -17,8 +17,19 @@ turn can choose differently. Transport, replay, native Usage and billing retain
17
17
  the actual model identity. Auto never crosses providers/accounts or changes an
18
18
  explicit real-model selection.
19
19
 
20
- The virtual preference remains the Session's pending selection while actual
21
- request headers record real models. For an otherwise unselected Session that
20
+ The virtual preference normally remains the Session's pending selection while actual
21
+ request headers record real models. A strict plugin-owned public Session projection
22
+ also retains the latest owner-explicit native `model/selection`: if a virtual
23
+ request header consumes native pending, later concrete headers cannot erase that
24
+ Auto intent. Cold refolding restores intent, not old decisions or reasons.
25
+ Copied fork/child prefixes alone do not establish an owner-explicit choice;
26
+ native pending and the existing parent-following rules remain independent.
27
+ Missing or malformed continuity evidence produces a named diagnostic instead of
28
+ guessing Auto from a real model. A new explicit fixed or other-provider selection
29
+ replaces intent for future turns, even after its native pending is consumed.
30
+ Already admitted Auto turns retain their real route through late picker changes.
31
+ This projection neither writes history/settings nor changes native projection ownership.
32
+ For an otherwise unselected Session that
22
33
  inherits default Auto, the first request persists the exact virtual
23
34
  `model/selection` intent. A later explicit selection wins. Existing fixed
24
35
  Sessions are not converted because the global default changes.
package/lib/client.js CHANGED
@@ -6453,7 +6453,7 @@ Set the \`cycles\` parameter to \`"ref"\` to resolve cyclical schemas with defs.
6453
6453
  "insufficient-evidence": zh ? "本地规则尚不能确定任务需求" : "local rules cannot establish task demand"
6454
6454
  };
6455
6455
  const 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]}.` : explanation.assessment.diagnostic === "timeout" && explanation.assessment.demand === "unknown" ? zh ? `任务难度判断超时,需求仍不确定;按当前偏好兜底选择 ${category[explanation.selectedCategory]}。` : `Task assessment timed out; demand remains uncertain. The preference policy selected ${category[explanation.selectedCategory]} as a fallback.` : zh ? `本轮判断为${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]}.`;
6456
- const 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.";
6456
+ const choice = explanation.method === "continuity" ? zh ? `该分类有 ${explanation.categoryCandidateCount} 个可容纳候选;上一轮模型仍合格,因此按连续性策略保留。这不是语义排名,也不证明它最适合任务。` : `This category has ${explanation.categoryCandidateCount} fitting candidates. Kept the previous model because it remains eligible, prioritizing continuity; this is not a semantic ranking or proof it is best for the task.` : 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 ${explanation.categoryCandidateCount} fitting category candidates by stable equal-weight allocation, not a quality ranking.` : zh ? "所有候选均不适配;这不是成功的偏好匹配。" : "All candidates are oversized; this is not a successful preference match.";
6457
6457
  const diagnostic = explanation.assessment.diagnostic === void 0 ? void 0 : explanation.assessment.diagnostic === "disabled" ? zh ? "语义判断实验未开启;这是本地规则的判断边界,不代表会话没有上下文。" : "Semantic assessment is disabled; this is a limit of local rules, not proof that the conversation lacks context." : explanation.assessment.diagnostic === "timeout" ? zh ? "辅助判断未在等待预算内完成,不代表正式回答超时或任务简单;本轮保留需求不确定,再按偏好策略选模。" : "The auxiliary assessment did not complete within its waiting budget. This is not a timeout of the main answer or evidence of simplicity; selection used the unknown-demand preference policy." : zh ? `语义判断未提供完整可用依据(${{
6458
6458
  unavailable: "没有可用分类模型",
6459
6459
  "invalid-result": "结果无效",