dsh-github-copilot 0.4.0-alpha.100

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.
Files changed (156) hide show
  1. package/AGENTS.md +312 -0
  2. package/CHANGELOG.md +479 -0
  3. package/CONTRIBUTING.md +46 -0
  4. package/LICENSE +21 -0
  5. package/README.md +203 -0
  6. package/README.zh.md +202 -0
  7. package/SECURITY.md +29 -0
  8. package/cordis.patch.yml +16 -0
  9. package/deployment-baseline.json +1299 -0
  10. package/docs/agent-readiness.md +110 -0
  11. package/docs/auto-task-routing.md +234 -0
  12. package/docs/automatic-model-routing.md +205 -0
  13. package/docs/copilot-accounts.md +236 -0
  14. package/docs/copilot-compaction.md +120 -0
  15. package/docs/copilot-usage.md +125 -0
  16. package/docs/dual-model.md +66 -0
  17. package/docs/image-input-compatibility.md +64 -0
  18. package/docs/images/copilot-accounts-credits.png +0 -0
  19. package/docs/images/copilot-accounts.png +0 -0
  20. package/docs/images/copilot-model-preferences.png +0 -0
  21. package/docs/images/copilot-search-routing.png +0 -0
  22. package/docs/images/dual-model-desktop.png +0 -0
  23. package/docs/images/dual-model-mobile.png +0 -0
  24. package/docs/images/dual-model-provenance.json +35 -0
  25. package/docs/images/parent-model-follow.png +0 -0
  26. package/docs/manual-compaction-recovery.md +64 -0
  27. package/docs/model-compatibility-acceptance.md +164 -0
  28. package/docs/npm-distribution.md +149 -0
  29. package/docs/official-first-016-alpha2.md +66 -0
  30. package/docs/official-first-020-rc1.md +32 -0
  31. package/docs/official-first-020-rc2.md +16 -0
  32. package/docs/session-search-routing.md +133 -0
  33. package/docs/single-route-migration.md +130 -0
  34. package/docs/web-lifecycle-rc2.md +55 -0
  35. package/lib/client.js +10822 -0
  36. package/lib/client.js.map +1 -0
  37. package/lib/copilot-identity-DRZY_d2X.js +15 -0
  38. package/lib/index.js +11689 -0
  39. package/lib/manual-compaction-recovery.js +315 -0
  40. package/lib/remote.js +734 -0
  41. package/lib/request-budget-DLY_aRq0.js +87 -0
  42. package/lib/routed-web.js +120 -0
  43. package/lib/search-routing-CqysPLve.js +792 -0
  44. package/lib/types/account-model-auth.d.ts +15 -0
  45. package/lib/types/account-model-catalog.d.ts +70 -0
  46. package/lib/types/account-model-source.d.ts +85 -0
  47. package/lib/types/agent-compaction.d.ts +4 -0
  48. package/lib/types/authorization-controller.d.ts +119 -0
  49. package/lib/types/auto-assessment-evidence.d.ts +46 -0
  50. package/lib/types/auto-model-host.d.ts +27 -0
  51. package/lib/types/auto-model-intent.d.ts +36 -0
  52. package/lib/types/auto-model-presentation.d.ts +57 -0
  53. package/lib/types/auto-model-routing.d.ts +59 -0
  54. package/lib/types/auto-selection-explanation.d.ts +9 -0
  55. package/lib/types/auto-task-assessment.d.ts +47 -0
  56. package/lib/types/auto-task-classifier.d.ts +9 -0
  57. package/lib/types/background-compaction.d.ts +30 -0
  58. package/lib/types/client.d.ts +115 -0
  59. package/lib/types/compact-account.d.ts +42 -0
  60. package/lib/types/compaction-pressure.d.ts +23 -0
  61. package/lib/types/compatibility.d.ts +17 -0
  62. package/lib/types/config.d.ts +89 -0
  63. package/lib/types/content-file.d.ts +10 -0
  64. package/lib/types/context-evidence-ui.d.ts +11 -0
  65. package/lib/types/context-evidence.d.ts +108 -0
  66. package/lib/types/context-usage.d.ts +4 -0
  67. package/lib/types/copilot-account-presentation.d.ts +7 -0
  68. package/lib/types/copilot-accounts-card.d.ts +30 -0
  69. package/lib/types/copilot-accounts-host.d.ts +75 -0
  70. package/lib/types/copilot-accounts-remote.d.ts +77 -0
  71. package/lib/types/copilot-accounts-types.d.ts +42 -0
  72. package/lib/types/copilot-auth.d.ts +25 -0
  73. package/lib/types/copilot-grant.d.ts +11 -0
  74. package/lib/types/copilot-identity.d.ts +11 -0
  75. package/lib/types/copilot-request.d.ts +11 -0
  76. package/lib/types/copilot-stream-liveness.d.ts +23 -0
  77. package/lib/types/copilot-usage-card.d.ts +30 -0
  78. package/lib/types/copilot-usage-host.d.ts +58 -0
  79. package/lib/types/copilot-usage-normalize.d.ts +9 -0
  80. package/lib/types/copilot-usage-remote.d.ts +63 -0
  81. package/lib/types/copilot-usage-types.d.ts +20 -0
  82. package/lib/types/copilot-usage-ui.d.ts +4 -0
  83. package/lib/types/current-provider.d.ts +59 -0
  84. package/lib/types/deepseek-search-fallback.d.ts +11 -0
  85. package/lib/types/dual-model-host.d.ts +119 -0
  86. package/lib/types/dual-model-remote.d.ts +63 -0
  87. package/lib/types/dual-model-types.d.ts +34 -0
  88. package/lib/types/external-link.d.ts +3 -0
  89. package/lib/types/failure.d.ts +38 -0
  90. package/lib/types/http.d.ts +48 -0
  91. package/lib/types/image-input-admission.d.ts +5 -0
  92. package/lib/types/index.d.ts +39 -0
  93. package/lib/types/manual-compaction-recovery.d.ts +52 -0
  94. package/lib/types/migration-status.d.ts +50 -0
  95. package/lib/types/model-exclusions.d.ts +15 -0
  96. package/lib/types/model-preferences-card.d.ts +9 -0
  97. package/lib/types/model-protocol.d.ts +31 -0
  98. package/lib/types/native-select-style.d.ts +4 -0
  99. package/lib/types/parent-model-follow-card.d.ts +41 -0
  100. package/lib/types/parent-model-follow.d.ts +44 -0
  101. package/lib/types/pi-provider-bridge.d.ts +22 -0
  102. package/lib/types/plan.d.ts +148 -0
  103. package/lib/types/pre-step-pressure.d.ts +8 -0
  104. package/lib/types/preview-provider.d.ts +66 -0
  105. package/lib/types/preview-route.d.ts +74 -0
  106. package/lib/types/probe.d.ts +39 -0
  107. package/lib/types/reasoning-presentation.d.ts +55 -0
  108. package/lib/types/remote-codec.d.ts +10 -0
  109. package/lib/types/remote.d.ts +199 -0
  110. package/lib/types/replay-recovery-host.d.ts +28 -0
  111. package/lib/types/replay-recovery-remote.d.ts +14 -0
  112. package/lib/types/replay-recovery-types.d.ts +24 -0
  113. package/lib/types/replay-recovery-ui.d.ts +26 -0
  114. package/lib/types/replay-recovery.d.ts +17 -0
  115. package/lib/types/request-body-evidence.d.ts +34 -0
  116. package/lib/types/request-body-timeout-marker.d.ts +2 -0
  117. package/lib/types/request-body-timeout.d.ts +4 -0
  118. package/lib/types/request-budget.d.ts +58 -0
  119. package/lib/types/request-upload-evidence.d.ts +23 -0
  120. package/lib/types/response-error-body.d.ts +3 -0
  121. package/lib/types/responses-reasoning-text.d.ts +24 -0
  122. package/lib/types/responses-reasoning.d.ts +7 -0
  123. package/lib/types/responses-replay-compat.d.ts +30 -0
  124. package/lib/types/route-ownership.d.ts +70 -0
  125. package/lib/types/routed-web.d.ts +50 -0
  126. package/lib/types/search-backend.d.ts +16 -0
  127. package/lib/types/search-routing-host.d.ts +19 -0
  128. package/lib/types/search-routing-policy.d.ts +18 -0
  129. package/lib/types/search-routing-remote.d.ts +20 -0
  130. package/lib/types/search-routing.d.ts +43 -0
  131. package/lib/types/search-tool-routing.d.ts +22 -0
  132. package/lib/types/serialize.d.ts +106 -0
  133. package/lib/types/settings-reader.d.ts +16 -0
  134. package/lib/types/sse.d.ts +28 -0
  135. package/lib/types/temporary-models.d.ts +25 -0
  136. package/lib/types/tool-schema-compat.d.ts +17 -0
  137. package/lib/types/traditional-search.d.ts +17 -0
  138. package/lib/types/turn-model-provenance.d.ts +45 -0
  139. package/lib/types/turn-selection-card.d.ts +10 -0
  140. package/lib/types/turn-selection-host.d.ts +11 -0
  141. package/lib/types/turn-selection-remote.d.ts +13 -0
  142. package/lib/types/turn-selection.d.ts +119 -0
  143. package/lib/types/turn-usage-evidence.d.ts +66 -0
  144. package/lib/types/turn-usage-notice.d.ts +7 -0
  145. package/lib/types/types.d.ts +69 -0
  146. package/lib/types/usage.d.ts +26 -0
  147. package/lib/types/watchdog.d.ts +22 -0
  148. package/lib/types/web-delegate.d.ts +12 -0
  149. package/lib/types/web-search-routing-card.d.ts +16 -0
  150. package/lib/types/web-search-routing-config.d.ts +12 -0
  151. package/lib/types/wire-anthropic.d.ts +37 -0
  152. package/lib/types/wire.d.ts +56 -0
  153. package/lib/web-delegate.js +9 -0
  154. package/output/auto-model-mockup/index.html +207 -0
  155. package/package.json +236 -0
  156. package/scripts/check-search-composition.mjs +191 -0
package/CHANGELOG.md ADDED
@@ -0,0 +1,479 @@
1
+ # Changelog
2
+
3
+ ## 0.4.0-alpha.100 (prepared)
4
+
5
+ - Add independently authorized GitHub Copilot accounts through official OAuth and the existing DSH credentials service (#297). Preserve the canonical compatibility account; inactive-account refresh remains lazy and record-bound.
6
+ - Manage, confirm switches, remove inactive authorizations and reauthorize the same identity in Models. Managed-only evidence, target preflight, activity fencing and path-level selector CAS prevent unsafe switches; no history, selected-model or default rewrites.
7
+ - Show the current GitHub identity read-only in Credits details, paired with account-bound quota evidence. Clear mounted readings during account changes and reject late responses; quota snapshots are not proof of cross-app AI-credit aggregation.
8
+ - Preserve existing Remotes and add a separate strict account namespace. Native gateway, adapter and synthetic on-disk restart tests complement the required full CI; no live OAuth, installation or loaded Desktop success is implied.
9
+
10
+ ## 0.4.0-alpha.99 (prepared)
11
+
12
+ - 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.
13
+ - 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.
14
+ - 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.
15
+
16
+ ## 0.4.0-alpha.98 (prepared)
17
+
18
+ - Retain owner-explicit Auto intent through a strict public Session projection when a native virtual request header consumes pending selection (#293).
19
+ - 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.
20
+ - 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.
21
+
22
+ ## 0.4.0-alpha.97 (prepared)
23
+
24
+ - 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).
25
+ - Keep unavailable, ambiguous and bounded-work evidence explicit; dispose subscriptions when Fetch settles. Local submission is not upload duration, kernel ACK or supplier receipt.
26
+ - Preserve Fetch arguments, Response/errors, payload/replay/history, connection/proxy/dispatcher and native retry behavior. This supplies future failure evidence, not a demonstrated supplier 408 cure.
27
+
28
+ ## 0.4.0-alpha.96 (prepared)
29
+
30
+ - Surface verified replay-recovery evidence automatically when an eligible Session opens or its native turn settles; ordinary conversations remain uncluttered (#289).
31
+ - Add explicit next-matching-turn or bounded Session consent, defaulting the choice to one turn. All native steps/retries share that turn's authorization, which is consumed at turn end. Confirmation never sends or retries a message.
32
+ - Preserve exact-item/proof guards, old Remote identities and native history ownership. Add native gateway and desktop/mobile component coverage, local dismissal, explicit read failures and expiry updates. This is not a supplier 408 fix.
33
+
34
+ ## 0.4.0-alpha.95 (prepared)
35
+
36
+ - Preserve native usage and chunk order before restoring dispatch-local byte-idle or replay errors (#285). Previously the owned error could be thrown before the SDK's terminal usage, discarding both failed-zero and real nonzero samples.
37
+ - Prove the correction with unchanged native Responses and Anthropic timeout regressions. Preserve structured errors, cancellation, retry policy and wire/history ownership.
38
+ - Document the separate native failed-zero context-pressure limitation and retry backoff timing. This is an accounting correction, not a cure for supplier HTTP 408 or a replacement native context meter; loaded Desktop acceptance remains unverified.
39
+
40
+ ## 0.4.0-alpha.94 (prepared)
41
+
42
+ - Count native Responses tool-output image spans under `function_call_output.output[]` in bounded timeout diagnostics (#287), rather than misclassifying them as remaining history.
43
+ - Preserve unknown output shapes, exact original byte spans, total-byte partitions, native image budgets, history/replay, failure classification and transport. Old stored diagnostics are not rewritten; this corrects evidence, not recurring supplier 408s.
44
+
45
+ ## 0.4.0-alpha.93 (prepared)
46
+
47
+ - Add default-off, explicit session replay recovery after verified Responses scope rejection. Confirming the disclosed loss omits only matching old encrypted items and their summaries on later requests; stored history, tool pairs and new reasoning remain unchanged.
48
+ - Bind recovery to native Agent lookup, initiating request signals, model/account proof and bounded Host-lifetime evidence. Do not automatically retry or treat this as a 408 fix.
49
+
50
+ ## 0.4.0-alpha.92 (prepared)
51
+
52
+ - Match the historical context disclosure title to neighboring composer statistics using the existing secondary font-size/line-height tokens, normal weight and tertiary label color (#283).
53
+ - Preserve native disclosure interaction, localization, historical evidence, context metering and Usage ownership; this changes presentation only.
54
+
55
+ ## 0.4.0-alpha.91 (prepared)
56
+
57
+ - Prevent known continuing-step managed input pressure before native model-attempt admission through the public `agent/pre-step` seam (#280), using the initiating Agent's current compaction service and one native transaction.
58
+ - Require a committed owned route and a proven absent pending model change. Respect disabled auto, overflow policy, cancellation and native failures; retain final converted admission and stream-pressure fallback.
59
+ - Prove the before/after accounting boundary with the unchanged rc.2 AgentLoop: successful early reduction removes the unsampled pressure attempt and preserves complete native two-step Usage. First-step/pending selections, new prefix growth and final hard refusals remain limitations; no historical totals, usage samples, retry counters or model selections are rewritten.
60
+
61
+ ## 0.4.0-alpha.88 (prepared)
62
+
63
+ - Capture bounded classifier model, total elapsed/budget, adapter-start, first-text, native finish and validation milestones in immutable ephemeral Auto turn evidence (#272). No prompts, raw outputs, credentials, replay or usage are recorded.
64
+ - Reduce auxiliary classification to 128 output tokens with compact demand/fixed-signal JSON. Request off reasoning only when supplier metadata and the published prepared-model capabilities agree; unsupported controls retain native policy.
65
+ - Explain semantic timeout as unknown-demand preference fallback and expose auxiliary milestones inside the existing bilingual disclosure. Preserve the 8s end-to-end deadline, cancellation/revocation, omitted-context guards and single attempt; no longer timeout, retries, guessed model-health ranking or measured improvement claim.
66
+
67
+ ## 0.4.0-alpha.87 (prepared)
68
+
69
+ - Remove managed terminal-zero usage filtering so native failure/cancellation samples remain available to Core's shared turn and cumulative accounting (#270).
70
+ - Retain unreliable failed-zero classification only in the independent historical context disclosure. Native zero pressure remains a known limitation, not a reason to remove shared usage or invent supplier receipts.
71
+ - Add native Responses zero/Anthropic nonzero cancellation and unchanged Core whole-turn regressions. Preserve errors, cancellation, retry and old histories; genuinely missing usage/lifecycle and timeout limitations remain explicit.
72
+
73
+ ## 0.4.0-alpha.86 (prepared)
74
+
75
+ - Enable bounded semantic Auto assessment by default for locally unknown tasks at the user's explicit request (#267). Preserve existing explicit false settings and document the opt-out.
76
+ - Apply the same default to unset managed-route settings, while retaining local known-task short-circuiting, one auxiliary call per admitted turn, strict output, timeout/cancellation/account guards and omitted-context protection.
77
+ - Disclose additional supplier charges outside native Chat Usage. Default rollout does not claim labeled-corpus calibration, change native history or perform profile installation/restart.
78
+
79
+ ## 0.4.0-alpha.84 (prepared)
80
+
81
+ - Add bounded numeric composition of the actual final request for verified Copilot `408 / user_request_timeout`, distinguishing conversation, tool definitions, protocol-owned system/instructions and residual framing (#263).
82
+ - Within conversation, count structural image blocks and opaque replay as disjoint wire-byte subsets without decoding or reporting their content. Missing, unsupported, malformed and work-limited evidence remains explicit.
83
+ - Capture per-dispatch fetch-to-response-header elapsed time before clone observation, honestly labeled as round trip rather than upload duration. Preserve native payload/response bytes, failure classification, retry, cancellation and credentials; no automatic settings change, trimming, compaction or model switch. Diagnostics guide explicit mitigation, not a claim to cure recurring 408.
84
+
85
+ ## 0.4.0-alpha.83 (prepared)
86
+
87
+ - Reduce avoidable Auto uncertainty through bounded difficult-task continuation and isolated fenced mechanical-transform recognition (#262); preserve unknown for new, incomplete and ambiguous work.
88
+ - Prioritize current and preceding user requests in the bounded semantic projection, preserving complete JSON rows, chronological order and omitted-context downshift protection.
89
+ - 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.
90
+
91
+ ## 0.4.0-alpha.82 (prepared)
92
+
93
+ - 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).
94
+ - Distinguish the exact recorded local pre-dispatch input-budget block, other settlements without usage, and incomplete history. A recovered turn can contain reported successful steps without a provable whole-turn total.
95
+ - Preserve native Usage when complete; do not fabricate zero usage, aggregate partial totals, change history or replace compaction/retry ownership. Record the observed plugin/Core integration gap and its limits.
96
+
97
+ ## 0.4.0-alpha.81 (prepared)
98
+
99
+ - Route Auto through current supplier Powerful/Versatile/Lightweight categories and contextual task evidence instead of advertised-capacity bands (#258). Preserve hard input/image/exclusion/account admission and native no-fit recovery.
100
+ - Use Lightweight for isolated greetings even with Intelligence, retain capable targets for short hard work, prefer suitable previous-model continuity, and distribute only equal-category candidates.
101
+ - Explain the actual captured turn choice, category fallback and continuity/tie-break in the footer; old records explicitly lack detailed reasons rather than displaying generic reconstructed text.
102
+ - Add an off-by-default bounded semantic-assessment experiment through the existing native adapter. Strict output, omitted-context conservatism, cancellation and diagnostic fallback remain; extra calls add latency and supplier charges outside native Chat Usage. Synthetic tests do not establish calibrated accuracy or GitHub-private parity.
103
+
104
+ ## 0.4.0-alpha.80 (prepared)
105
+
106
+ - Observe real HTTP/SSE byte progress on managed requests, retaining the original five-minute byte-idle bound while allowing at most ten minutes without a native assistant chunk (#253).
107
+ - Preserve native SDK parsing, exact stream bytes, caller cancellation and retry ownership; endless heartbeats remain bounded and explicit WebSocket/auto transport keeps native behavior.
108
+ - Add separate managed-chat timeout/opt-out settings and a configurable native image-request budget with the unchanged 20 MiB default. No hidden history trimming, automatic model switch, extra retry or claim to cure HTTP 408.
109
+ - Reproduce heartbeat-only semantic timeouts and verify real late output, actual byte stalls and cleanup through unchanged native adapters.
110
+
111
+ ## 0.4.0-alpha.79 (prepared)
112
+
113
+ - Bind retained turn selection reads through the public `TypertRemoteService` and `@Remote get`, fixing the Host registration missing from the alpha.76 Client-only correction (#254).
114
+ - Cover actual Client-to-Host calls in source-discovery and strict modes, including Auto/Manual, missing evidence, missing/denied identities and unload; preserve lookup checks and the existing bounded decision store.
115
+
116
+ ## 0.4.0-alpha.78 (prepared)
117
+
118
+ - Optimize individual Exclude/Restore saves with a strict preferences-only Remote response, avoiding credential/account-status reads and live-session scans (#252). Keep single-row immediate persistence, CAS conflicts, invalidation and failure recovery; no batch UI.
119
+ - Permit exclusion of selected models without rewriting Session/default selections. Preserve already admitted native turns through tool steps and retries, reject excluded models on new turns, and retain account/token/metadata/cancellation guards.
120
+ - Native configuration serialization, file locks and Loader synchronization still gate save completion; no live latency threshold or loaded Desktop upgrade is claimed.
121
+
122
+ ## 0.4.0-alpha.77 (prepared)
123
+
124
+ - Resolve automatic compaction pressure and Auto recovery availability from the initiating Agent's preset through the existing public service lookup (#248), including Desktop's isolated compaction groups.
125
+ - Never borrow global recovery when a bound preset has no engine; preserve disabled/zero-retry policy, native transactions, cancellation, and manual-only segmented recovery. No new command, timeout reclassification, Core change, or automatic model switch.
126
+ - Cover concurrent enabled/disabled/absent preset engines with unchanged official rc.2 AgentLoop, durable summary/rebuilt-request and cancellation evidence; add the frozen preset dependency closure to both required CI and release gates.
127
+
128
+ ## 0.4.0-alpha.76 (prepared)
129
+
130
+ - Fix fresh master/lead Auto footer reads rejected by the native Client gateway when an ambient agent context reduced the explicit two-argument call to one (#249). Preserve native Host lookup and strict codecs; do not substitute the ambient Session.
131
+ - Distinguish failed selection reads from successfully read missing evidence. Add a bounded, explicit same-turn Retry without rerunning inference or leaking raw errors.
132
+ - Reproduce the old descriptor failure and verify bound/unbound calls through the unchanged pinned Client gateway.
133
+
134
+ ## 0.4.0-alpha.75 (prepared)
135
+
136
+ - Add safe, actionable managed `COPILOT_REQUEST_BODY_TIMEOUT` guidance for the exact observed HTTP 408 request-body timeout, including final JSON UTF-8 bytes when observable (#246).
137
+ - Preserve native failure classification, retry metadata, payloads, cancellation and credential ownership; no new retry, timeout change, automatic compaction, hidden trimming or model switch.
138
+ - Share the existing bounded clone-only response observer with replay-scope classification and cover all three native HTTP protocols, unknown/oversized/stalled replies and concurrent dispatch isolation. This is diagnostic/recovery guidance, not a claim to eliminate upstream timeouts.
139
+
140
+ ## 0.4.0-alpha.74 (prepared)
141
+
142
+ - Add `/copilot-compact [status|cancel]` to the explicitly selected recovery engine when native commands/jobs are available (#244).
143
+ - Acknowledge a Session-owned background job promptly instead of holding Desktop's unary command request past its response-header timeout. Preserve native `/compact`, maintenance locks and compaction transactions.
144
+ - Deduplicate active starts, cancel and drain on teardown, and collect native settlement without waking the model or resuming a Goal. Admission is not completion; failed summaries retain their native diagnostics.
145
+
146
+ ## 0.4.0-alpha.73 (prepared)
147
+
148
+ - Prevent managed Copilot failure/cancellation terminal zero-usage samples from resetting native context pressure; preserve successful zero and real nonzero usage (#242).
149
+ - Add a strict plugin-owned historical context projection and an additive composer disclosure for replayed failed-zero readings. Last input-plus-cache samples are historical evidence, never current occupancy or inferred percentages.
150
+ - Revoke old samples on route changes, compaction, surface replacement and invalid attribution. Keep the Core meter, account Credits, durable history, retries and compaction ownership unchanged.
151
+
152
+ ## 0.4.0-alpha.72 (prepared)
153
+
154
+ - Retain the mounted account controller and model metadata across exclusion-driven Models rerenders by capturing the native traced Copilot Remote once per UI registration (#240).
155
+ - Avoid redundant account status and model-discovery work on provider/footer/section rerenders; preserve real registration teardown, credential invalidation, selection locks and exclusion CAS.
156
+ - Add regressions with fresh Remote proxies and a blocked second status read, plus exact published-gateway namespace identity evidence. No Core changes or Desktop install/restart.
157
+
158
+ ## 0.4.0-alpha.71 (prepared)
159
+
160
+ - Decouple the Auto selection footer from optional conversation projections and wait for its exact Remote namespace (#238).
161
+ - Use native completed-turn evidence to show retained Auto/Manual records when projection data is unavailable; keep incomplete attribution explicit without guessing historical selections or changing native Usage.
162
+ - Cover cold/incremental native assembly and session-scoped rendering, projection failures, pending dependencies and invalid completion evidence. Live Desktop recovery still requires loaded-version verification.
163
+
164
+ ## 0.4.0-alpha.70 (prepared)
165
+
166
+ - Require image-capable Auto candidates for retained historical user/tool images after text-only continuations; honor native image offload markers (#236).
167
+ - Validate actual native-projected image MIME against explicit account format evidence before managed dispatch, with request-local diagnostics and no image rewriting, model switch or additional retry.
168
+ - Document why durable attachment formats and filenames cannot prove outgoing MIME, and retain the original service-side rejection as unverified rather than declaring Grok text-only.
169
+
170
+ ## 0.4.0-alpha.69 (prepared)
171
+
172
+ - Add an explicit, opt-in manual compaction engine for compressible already-oversized managed Copilot histories. It uses a bounded public summary hook and native transactional replacement without changing the default compaction service. Include account-proof and failure boundaries in the recovery design.
173
+
174
+ ## 0.4.0-alpha.68 (prepared)
175
+
176
+ - Add one **Follow parent model** control in plugin settings. Enable once for existing and new supported Copilot subagents and Team mates; no Session IDs or per-child setup (#234).
177
+ - Preserve active turns, child-owned manual selections and independent child Auto decisions. Keep legacy explicit bindings without hidden migration.
178
+ - Record requirements, implementation plan and a synthetic screenshot of the actual settings component; use narrow settings CAS and preserve unsaved search drafts.
179
+
180
+ ## 0.4.0-alpha.66 (prepared)
181
+
182
+ - Add explicit native child/direct-parent model-follow bindings, disabled by default. Enrolled children follow fixed selections on their next turn or independently resolve the parent's Auto preference against their own context (#229).
183
+ - Preserve in-flight turn routing, explicit child selections, native histories, defaults, credentials and permissions. Unknown lineage, unsupported children and missing parent evidence fail with named diagnostics rather than guessed models.
184
+ - Record exact-source inheritance research, configuration boundaries and validation limits. Native creation labels are not rewritten, and existing children are never automatically enrolled.
185
+
186
+ ## 0.4.0-alpha.65 (prepared)
187
+
188
+ - Fix Auto input undercount by using Core's public TokenMeter instead of passing Core messages to pi-ai's incompatible estimator and swallowing its errors (#230). Include assistant reasoning/tool history and retain native current-surface/tool-envelope measurement as a conservative floor.
189
+ - Fail Auto explicitly when native measurement is missing or invalid; fixed model selection remains available. Recognize native compact-checkpoint provenance without changing messages or same-turn routing.
190
+ - This is a preventive input-fit fix, not recovery of already oversized manual summaries. The independent summary-input overflow investigation remains open in #228; no chunking, alternate summarizer, history deletion, or extra retry loop is introduced.
191
+
192
+ ## 0.4.0-alpha.64 (prepared)
193
+
194
+ - Omit only explicitly empty completed reasoning shells from managed Responses outgoing payloads, without changing durable history, opaque encrypted content, public summaries or tool call/result pairing (#226).
195
+ - Preserve fail-closed handling for partial, unknown and reference-dependent replay. Existing HTTP-408 retry snapshots may reuse the same omitted shell only when the entire original request and retry scope match; no new retry loop or model switch is introduced.
196
+
197
+ ## 0.4.0-alpha.63 (prepared)
198
+
199
+ - Restore the completed-reply Auto preference and recorded-reason disclosure on official Desktop by reading ChatNodeStore's public iterable `values()` instead of requiring a JavaScript `Map` (#222). Retain Map support and diagnose an unavailable collection once without querying another turn.
200
+ - Preserve exact Session/turn lookup, stale-response isolation, unknown/manual evidence rules, native Usage model display, and existing selection policy. No history migration or Core changes are required.
201
+
202
+ ## 0.4.0-alpha.62 (prepared)
203
+
204
+ - Preserve model preferences through Client authorization decoding and make hidden exclusions a native live Config field, fixing both invisible controls and rejected settings writes (#223).
205
+ - Add All/Enabled/Excluded filtering, searchable model rows, explicit selected-model locks, status-only Retry and named read-only diagnostics. Keep discovery details separate and reject stale action results after account changes or unmount.
206
+ - Add shared-account-to-Manage regression coverage and unchanged native SettingsForms CAS/fiber/persistence evidence. No Core, credentials, histories or native picker layout are changed.
207
+
208
+ ## 0.4.0-alpha.61 (released)
209
+
210
+ - Preserve the exact normalized managed Responses request across Core's freshly prepared retry attempts only after the previous request actually received HTTP 408. Short-lived state is scoped to the original turn signal, Session, model and account proof, and cleared on new step/turn end. Retry-only item references or incomplete ID-bearing items must identify matching complete prior items without contradictory fields, and all other payload and transcript bytes must agree; changed inputs, concurrent requests and cold resumes fail closed (#217).
211
+ - Keep native retry policy, SDK transport, durable replay, credentials, and model selection unchanged. A synthetic native-SDK regression covers two 408 responses followed by a reference-only third attempt; real account acceptance and Desktop installation remain unverified.
212
+
213
+ ## 0.4.0-alpha.60 (released)
214
+
215
+ - Add exact account-model exclusions under **GitHub Copilot → Manage → Model preferences**, with local search, visible/excluded counts, explicit **Exclude/Restore** actions, and retained rows for excluded IDs temporarily absent from account metadata (#192).
216
+ - Remove excluded IDs from the managed picker, every Auto candidate pool, aggregate Auto modalities, hosted-search route facts, and request-budget admission. Stale picker or direct requests fail with `COPILOT_PREVIEW_MODEL_EXCLUDED`; Auto fails closed when no eligible model remains.
217
+ - Protect the current fixed managed selection using public default and live Session projection/request-header evidence. Exclusion writes use path-level settings CAS and never rewrite history, switch a Session, mutate the account grant, or change Core's flat picker layout.
218
+
219
+ ## 0.4.0-alpha.59 (released)
220
+
221
+ - Fix unavailable Web search routing on official rc.2 by exposing the plugin's live routing fields through native Config projection (#216). Retain narrow CAS writes, separate legacy override reset, ordinary safety configuration and live snapshot consumption without remounting.
222
+ - Replace the manually populated settings test double with a disposable native Loader/ConfigEditor/SettingsForms fixture covering projection, save, conflict rejection and restart persistence.
223
+
224
+ ## 0.4.0-alpha.58 (prepared)
225
+
226
+ - Add compact Auto preference and expandable recorded reasons after the native reply time through the public assistant-actions slot; explicit fixed selection shows Manual, absent evidence shows Selection unknown. Remove the separate Model details candidate and repeated model name; actual models stay in native Usage.
227
+ - Keep new selection evidence in a bounded Host-lifetime store with native agent-scoped lookup, not incompatible durable events. Restart, disposal and eviction lose this evidence; compatible historical Auto records remain readable. Native Usage aggregation and its missing-route behavior are unchanged.
228
+ - Preserve plugin-owned credential invalidation and disposal reasons as specific `ABORTED` diagnostics, without changing transport retries, timeouts or partial output.
229
+ - Add offline native-SDK regressions distinguishing HTTP 408, transport termination, missing Responses terminal events, true cancellation and zero-usage measurement limits. These changes do not claim to fix upstream timeouts, disconnects or native context-meter resets.
230
+
231
+ ## 0.4.0-alpha.57 (released)
232
+
233
+ - Keep plugin-owned rc.2 lifecycle guidance visible before disable, removal, or upgrade, because the plugin UI is unloaded before Desktop can report dependent Web entries pending on a service (#185).
234
+ - Document that an already-Off plugin with only `pending (waiting for service: …)` dependent entries requires a full Desktop restart rather than repeated toggling, without reclassifying other native failures.
235
+ - Include the bounded rc.2 Web lifecycle characterization in the published package.
236
+
237
+ ## 0.4.0-alpha.56 (released)
238
+ - Pin `undici` to 6.28.1 to resolve the observed 6.29.0 installation failure in the configured package mirror (#212). Keep the lockfile registry-neutral and preserve the quota-only TLS dispatcher; this does not establish full Desktop installation or activation.
239
+ ## 0.4.0-alpha.55 (prepared)
240
+
241
+ - Use a quota-only, reusable TLS-verified dispatcher with Node default and system CA roots for the GitHub Copilot usage endpoint (#209). This removes reliance on launcher inheritance of `NODE_USE_SYSTEM_CA` for quota requests, without changing Core, sign-in, model/search transport or global trust.
242
+ - Show an explicit unavailable diagnostic if system trust cannot be read or the dispatcher cannot be created; preserve injected fetch isolation and close the dispatcher on disposal.
243
+
244
+ ## 0.4.0-alpha.54 (prepared)
245
+
246
+ - Stop writing optional Auto attribution events that official DSH 0.2.0-rc.2 cannot mark ignorable, preventing subsequent Session history reads from failing (#204). Auto routing, durable virtual selection, and Core's actual model provenance are unchanged; new turns omit the plugin-specific attribution footer.
247
+ - Add an explicit source-checkout recovery command using the unchanged official v4 codec. It validates a detached repaired copy and preserves a byte-exact original backup, without replacing live histories or restarting Desktop.
248
+
249
+ ## 0.4.0-alpha.53 (prepared)
250
+
251
+ - Fix Auto model concentration and context budget overflow by evaluating input fit as a hard routing condition before preference and distributing soft preferences across candidate bands with deterministic turn seeding (#199).
252
+ - Exclude candidates whose hard input limit cannot admit estimated turn input, preventing `COPILOT_CONTEXT_BUDGET_EXCEEDED` failures.
253
+ - When no candidate fits, route to the largest input capacity model and trigger compaction pressure on compressible history.
254
+ - Distribute soft preferences across upper/lower/center candidate bands via deterministic in-memory hash of `sessionId:turn`, preventing top catalog models from monopolizing all Sessions while freezing selections across retries and steps within the same turn.
255
+
256
+ ## 0.4.0-alpha.52 (prepared)
257
+
258
+ - Place Copilot Auto model attribution after the settled turn clock on the same footer row, and reveal it only on hover or keyboard focus, matching usage and time chrome.
259
+
260
+ ## 0.4.0-alpha.51 (prepared)
261
+
262
+ - Fix web search routing settings save by reading and writing `searchRouting` under the existing `github-copilot` settings namespace, avoiding the unexported `installSection` limitation in official `@deepseek-ai/dsh-settings@0.2.0-rc.2` (#195).
263
+ - Coordinate CAS revisions between web search routing and legacy `searchModel` resets within the shared `github-copilot` namespace while preserving unsaved drafts.
264
+ - Remove fake separate namespace and `installSection` mocks from synthetic test fixtures and align contracts.
265
+
266
+ ## 0.4.0-alpha.50 (prepared)
267
+
268
+ - Add a visible restart banner in `WebSearchRoutingCard` when the routed search service composition is unready, guiding users to cold-restart Desktop (#196).
269
+ - Streamline the plugin details configuration page by removing the internal `HostedSearchSettingsCard` debugging form from `plugins.bundle.config`, leaving only `WebSearchRoutingCard`.
270
+ - Add guidance on `NODE_USE_SYSTEM_CA=1` for Copilot Quota TLS verification when using network accelerators or local proxies like Watt Toolkit (Steam++).
271
+ - Update documentation on cold-restart requirements for `cordis.patch.yml` service recomposition and Windows system CA trust.
272
+
273
+ ## 0.4.0-alpha.49 (prepared)
274
+
275
+ - Add bounded soft capacity preferences to Copilot Auto model routing (#192).
276
+
277
+ ## 0.4.0-alpha.48 (prepared)
278
+
279
+ - Keep the managed Copilot Auto model selected across turns when inherited as a new Session's default; record its durable intent before the first real-model request header and restore the virtual route on later turns (#189).
280
+ - Preserve explicit fixed-model selections, real request headers and per-turn Auto decisions without changing existing Sessions with a recorded concrete route.
281
+ - Supersede unpublished alpha.47 after its immutable pre-created tag failed the protected exact-annotation check; no alpha.47 GitHub Release or npm package was created.
282
+
283
+ ## 0.4.0-alpha.47 (publication blocked)
284
+
285
+ - Publication stopped after packing and checksum verification because the immutable pre-created tag's annotation did not satisfy the protected exact-message check.
286
+ - The alpha.47 tag remains unchanged and is not reused; its prepared Copilot Auto selection fix is carried by alpha.48.
287
+
288
+ ## 0.4.0-alpha.46 (prepared)
289
+
290
+ - Distinguish known TLS certificate-verification failures from other Copilot quota network errors without revealing provider errors or disabling certificate validation (#186).
291
+ - Keep unavailable quota details readable: omit empty large-value columns and budget placeholders, use an opaque popover background, and bound long values within the dialog.
292
+
293
+ ## 0.4.0-alpha.45 (prepared)
294
+
295
+ - Add a fixed, allowlisted latest-observed authorization milestone to failed sign-in diagnostics without inspecting or retaining rejected values, provider responses, credentials, prompts, notices, or nested causes (#178).
296
+ - Clear the ephemeral milestone on cancellation, success, sign-out, and each new attempt; distinguish observed request, prompt, and notice progress from an inferred failure cause.
297
+ - Document the exact rc.2 hot-removal boundary: removing the routing bundle after activation can strand the public Web service isolate until restart, while the native manager already marks existing-package upgrades as restart-required. Do not remove and re-add the package to bypass that boundary.
298
+ - Supersede unpublished alpha.44 after its immutable tag failed the protected annotation check; no alpha.44 GitHub Release or npm package was created.
299
+
300
+ ## 0.4.0-alpha.43 (prepared)
301
+
302
+ - Put Web search routing and Copilot hosted-search options on the plugin bundle's detail page through the public `plugins.bundle.config` slot (#177).
303
+ - Keep sign-in, account status and model refresh in Models; retain the routing card's Models/section fallback if the bundle slot is unavailable.
304
+ - Save routing and hosted-search options independently to their existing namespaces with revision checks, preserving unchanged allowlists and explicit legacy model overrides. No settings, profile or Session is migrated on page open.
305
+
306
+ ## 0.4.0-alpha.42 (prepared)
307
+
308
+ - Add a plugin-owned virtual **Auto** model to the managed GitHub Copilot route, resolving once per Core turn to a real account model while preserving explicit selections and truthful request provenance (#174).
309
+ - Filter image turns by account-verified image capability, keep search independently owned, and preserve native subagent concrete-route inheritance.
310
+ - Reuse the existing local request-pressure signal and Core transactional compaction so a smaller-context Auto choice can recover with the same real model and no oversized wire request.
311
+ - Record credential-free Auto decisions for a public turn-tail `Auto · actual model` disclosure with an inline explanation, and correct virtual-Auto model-switch guidance before it reaches the model.
312
+ - Document the plugin-only MVP and interactive multi-provider mockup. The deterministic capacity heuristic does not claim model quality, latency, price, health, or parity with GitHub's private Auto router.
313
+
314
+ ## 0.4.0-alpha.40 (prepared)
315
+
316
+ - Sanitize authorization failures into allowlisted begin-stage and post-auth route-repair diagnostics in the existing Remote error field; unknown or legacy arbitrary errors remain generic.
317
+ - Stop retaining or logging raw authorization errors, including message, URL, response body, token, device code, and nested cause content; render safe stage-specific next steps.
318
+ - Add regressions for cancellation, successful authorization, route-repair failure, nested causes, secret-bearing errors, arbitrary thrown values, and backward-compatible generic Client behavior (#172).
319
+ - Diagnostics identify only where the flow failed; they do not explain or resolve the underlying OAuth failure.
320
+
321
+ ## 0.4.0-alpha.39 (prepared)
322
+
323
+ - Qualify the exact official DSH and signed Windows Desktop `0.2.0-rc.2` peer graph; retain rc.1 as historical evidence.
324
+ - Align the managed native adapter with pi-ai `0.87.1` transcript contexts and preserve truthful estimated input admission.
325
+ - Compose ordered bundle patch arrays in the read-only installation preflight through the public rc.2 API, with malformed/ordering regressions.
326
+
327
+ ## 0.4.0-alpha.38
328
+
329
+ - Add official DSH `0.2.0-rc.1` as the current exact compatibility target while retaining the existing baseline pins; use the corresponding public Core packages in development and tagged-source CI.
330
+ - Adapt plugin settings reads/listeners to the current public `SettingsForms` descriptor/event API while retaining legacy Core compatibility. Materialize volatile pi-ai provider profiles through their public getter.
331
+ - Preserve current first-class tool messages, older nested tool results and Core-owned developer history semantics in the plugin's guarded serializers; retain recursive legacy file detection when the Core predicate does not recognize older nested content.
332
+ - Update Remote/Gateway and serialized agent lifecycle regressions to match the published 0.2 contracts, and keep provider routing writes isolated from the Copilot settings namespace.
333
+ - Record the pinned official-first API review and qualification boundaries in `docs/official-first-020-rc1.md`. Local tests do not qualify the packaged Windows Desktop runtime, live OAuth/provider calls, or loaded Desktop activation.
334
+
335
+ ## 0.4.0-alpha.37 (prepared)
336
+
337
+ - Recognize both observed uncoded Responses scope-401 messages, including `input item does not belong to this connection` without `ID`, so that request-local rejection does not retire shared OAuth proof or abort neighboring calls (#164).
338
+ - Retain exact message matching, consistent outer/nested envelopes, genuine-auth code/type precedence, HTTP-status and bounded-body checks; do not broaden classification to arbitrary connection text.
339
+ - Exercise both variants through the real native adapter for concurrent isolation, same-lease independent dispatches, cancellation, credential reuse and genuine-auth renewal. The fixture reproduces the defect without cron.
340
+ - Keep payload normalization, encrypted history, model selection and transport unchanged. This contains a classification defect; it does not establish why the remaining historical item is rejected or claim complete live-session recovery. No Core, cron, dependency or installed-profile change is included.
341
+
342
+ ## 0.4.0-alpha.36 (prepared)
343
+
344
+ - Add managed Responses wire-only replay normalization through the native SDK's public `onPayload` seam, preserving durable history, encrypted reasoning, message phases and tool-result pairing (#162). Reference-only or incomplete ID-bearing forms fail explicitly instead of silently discarding history.
345
+ - Recognize the exact structured input-item connection-scope HTTP 401 without retiring shared OAuth proof or aborting neighboring calls. Surface a sanitized, dispatch-local `INVALID_REQUEST`; genuine and unrecognized authentication failures retain existing recovery.
346
+ - Preserve native transport, caller callbacks, cancellation, credentials and model selection. No automatic request retry, alternate adapter, Core/SDK patch, canonical-route takeover or live-profile mutation is introduced.
347
+ - Add real-SDK/pinned-Core synthetic regressions for replay shape, caller callback composition, bounded response inspection, parallel isolation, cancellation and genuine authentication recovery. Real Copilot acceptance of existing conversations, deployment and activation remain separate acceptance steps.
348
+
349
+ ## 0.4.0-alpha.35 (prepared)
350
+
351
+ - Match the account-usage control to native secondary statistics typography, line height and pill spacing (#160).
352
+ - Keep its label on one line within the host's available width, with ellipsis and a full-label tooltip for constrained layouts; independent details, focus and account-wide billing semantics remain unchanged.
353
+ - Remove the permanently unavailable Session credits section. Hide missing, malformed, zero or elapsed reset metadata instead of displaying the Unix epoch; prefer a valid snapshot reset, then a valid account-level date, without inventing dates or discarding otherwise valid quota amounts.
354
+ - Continue using the existing additive composer dock. Shared-row placement and responsive wrapping belong to the host layout; no private DOM relocation, native statistics replacement or new Core API dependency is introduced by this plugin.
355
+ - Preserve alpha.34's Model roles retirement and compatibility-only historical Session behavior. This presentation update does not install, activate or restart a profile.
356
+
357
+ ## 0.4.0-alpha.34 (prepared, PR-only)
358
+
359
+ - Retire the Model roles / planner-executor settings, legacy Settings fallback, and dedicated-session creation UI (#158); ordinary Sessions and Core-owned subagents remain the intended workflow.
360
+ - Remove obsolete role Client components and browser fixtures, with negative registration and adjacent account/search/usage regressions.
361
+ - Keep existing dedicated-session policy replay and strict Remote contracts for compatibility; reject role configuration writes and new dedicated roots explicitly, retaining evidence-only recovery of matching existing requests.
362
+ - Do not migrate or delete saved settings, histories, credentials, model choices or global defaults. Native parent-to-child model rules remain separately pending Core work, not a capability supplied by this retirement.
363
+ - Version metadata prepares the required PR release gate only; merge, publication, installation and activation are not part of this PR-only request.
364
+
365
+ ## 0.4.0-alpha.33 (prepared)
366
+
367
+ - Fix the account-usage composer component to call the public `useSession(selector)` hook with a selector instead of assuming a no-argument snapshot API (#156).
368
+ - Preserve native Session ownership and select only the open/current/not-removed predicate; no Core patch, credential change, network-policy change or automatic restart is required.
369
+ - Add positive canonical/managed Copilot rendering and required-selector lifecycle regressions. Earlier signed-out absence tests do not establish that the eligible-session control can render.
370
+
371
+ ## 0.4.0-alpha.32 (prepared)
372
+
373
+ - Add an optional, Copilot-session-scoped composer usage control without replacing the native Context meter or changing other providers (#153).
374
+ - Read bounded account quota snapshots through the canonical Host-owned OAuth grant and a separate strict usage Remote namespace; preserve the existing authorization contracts.
375
+ - Distinguish AI credits, legacy premium requests, individual budgets, shared pools, unavailable data and last-known snapshots. Never invent remaining credits from missing metadata, token estimates or another session.
376
+ - Keep per-session credits explicitly unavailable until a supported public native accounting seam exposes complete provider-reported usage; no second model transport, Core patch, live-profile installation or restart is included.
377
+
378
+ ## 0.4.0-alpha.31 (prepared)
379
+
380
+ - Retire the exact current managed-route token proof after an observed model HTTP 401, preserving the native failure without replay, logout or model switching (#152).
381
+ - Let the next independent caller renew a rejected token through native `Models.getAuth()` and the serialized canonical credential store, even if its stored expiry is still in the future. Never persist a projected expiration or replace a newer sign-in.
382
+ - Bound recovery per account using the existing cooldown setting (five minutes by default, one-second floor); reject identical-token renewal and ignore late generations, 403, cancellation and error strings without an observed HTTP response.
383
+ - Cover sign-out/sign-in with retained conversation history, native HTTP transports, concurrent renewal, credential races and cooldown. Synthetic evidence does not prove the original endpoint rejection cause, live OAuth/model acceptance, WebSocket recovery, publication or Desktop activation.
384
+
385
+ ## 0.4.0-alpha.30 (prepared)
386
+
387
+ - Route verification links on recognized Desktop v1 hosts through the existing same-window external-navigation handoff instead of popup creation; retain new tabs on web hosts (#150).
388
+ - Display a selectable verification address for manual browser handoff without claiming that the system browser opened or authorization succeeded.
389
+ - Remove the Compatibility and existing configurations disclosure from Manage. Preserve legacy configuration diagnostics, explicit repair, migration documentation, credentials and session behavior.
390
+ - Cover Desktop/web targets, manual URL handoff and expanded Manage across account states. Synthetic Client evidence does not establish live Desktop activation or OAuth success.
391
+
392
+ ## 0.4.0-alpha.29 (prepared)
393
+
394
+ - Fix search settings drafts and pending saves being reset by parent renders: capture stable traced Settings/routing Remote faces once per registration (#148).
395
+ - Make ordinary Web search settings provider-only, rename the final default to Fallback provider, and save both routing choices in one namespace CAS without model discovery or a model prerequisite.
396
+ - Resolve independent Copilot search models from current account-owned Responses metadata in deterministic order, with at most three capability-probed candidates and one final query. Preserve allowlists, explicit overrides, account/owner proof, cancellation and fallback policy; never borrow a Chat/global default or hardcode model IDs.
397
+ - Preserve existing model overrides and offer a separate explicit reset to automatic selection. Classify safe save/conflict diagnostics without exposing raw Remote errors.
398
+ - Add real pinned Settings/Client-codec contracts, mounted React/Cordis lifecycle tests, and isolated built-browser before/after reproduction. These are not live OAuth/search, published-release or installed Desktop acceptance claims.
399
+
400
+ ## 0.4.0-alpha.28 (prepared)
401
+
402
+ - Enforce plugin-owned estimated independent prompt and combined input/output budgets on the managed Copilot route while preserving advertised context capacity and native transport (#146).
403
+ - Signal eligible exact-session loop pressure before model dispatch through the existing official bounded compaction/rebuild path; respect disabled automatic recovery, model-specific zero retries and cancellation.
404
+ - Prefer supported minimal/low reasoning only for compaction with no supplied or materialized effort, without silently rewriting the requested summary output cap or changing ordinary conversations.
405
+ - Preserve structured local admission failures through SDK wrapping, current system/tool accounting, concurrent-call isolation and distinct output-truncation outcomes. Add configurable safety/pressure/summary policy and regression coverage.
406
+ - Keep already oversized manual summaries as explicit failures; this release does not claim chunked recovery, silent history deletion, automatic model switching, Core changes or local Desktop activation.
407
+
408
+ ## 0.4.0-alpha.27 (prepared)
409
+
410
+ - Remove the separate Model roles workspace dropdown: role settings remain profile-global and saving does not require a workspace (#144).
411
+ - Show the current workspace read-only for explicit dedicated-session creation, observing public Client lists on both legacy selection and alpha.2 main-view ownership. Missing or ambiguous selection disables creation rather than choosing a first/recent workspace.
412
+ - Retain one traced Remote face per UI registration so navigation preserves drafts and uncertain-create UUID/workspace/revision across footer/fallback remounts. Ordinary and existing sessions remain unchanged.
413
+ - Add mounted Cordis/React and current-workspace regressions without changing Core, dependencies, Host codecs or session policy.
414
+
415
+ ## 0.4.0-alpha.26 (prepared)
416
+
417
+ - Fix low-contrast native dropdown options in dark mode for planning/execution models, workspaces and search providers (#142).
418
+ - Pair opaque application-theme surfaces with primary or secondary foreground tokens on both selects and options, with readable system-color fallbacks on hosts without those tokens.
419
+ - Preserve disabled/unavailable choices and all selection, CAS and session behavior. Add focused regressions and remove fixture-only option colors that masked the production bug; no Core or dependency changes are required.
420
+
421
+ ## 0.4.0-alpha.25 (candidate)
422
+
423
+ - Append the ninth exact official target `dsh-v0.1.6-alpha.2` at `ddefc45fbc7f8e46dd73185e68295696d1297887`, retaining all older pins and exact `0.1.2-rc.1` development dependencies.
424
+ - Supply strict Remote `create()` factories with a legacy `schema` bridge over the same parser; preserve endpoint contracts and validation rather than weakening to `src-json`.
425
+ - Correct dedicated executor projection admission to native `subagent/descriptor` v3. Native descriptors were already v3 in rc.1; the plugin's old v1 assumption was a bug, not an upstream format migration.
426
+ - Bump the plugin projection cache to `stateVersion: 2` to force refolding. Unknown/v1/v2 descriptor histories fail closed and remain unmodified; recovery requires a reviewed new child, never relabeling or fabricated conversion.
427
+ - Add strict-codec, descriptor/projection and exact-alpha.2 contract fixtures, plus an [official-first comparison and retirement plan](./docs/official-first-016-alpha2.md) for retained custom surfaces.
428
+ - Evidence remains limited: source markers, local rc.1-backed focused tests and fifteen scoped exact-source runtime tests passed (alpha.2 contracts 8, Remote 1 and Session-context 6; supplemental resolver with official TypeScript `6.0.3`, declared `mime-types@3.0.2` and `ws@8.21.0`, and shared Zod `^4.4.3`, no source/dependency patches). Full local `pnpm verify` passed: 1373 Vitest tests with 2 expected skips, 176 tooling tests, typechecks/build/package smoke; pack/tarball verification passed. Broad frozen dependency installation remains blocked by the configured mirror's `node-addon-require-builtin@0.1.6` HTTP 404. Full official-root-helper and CI qualification remain pending. This candidate is not a published-artifact, live Desktop, OAuth or model-call compatibility claim.
429
+
430
+ ## 0.4.0-alpha.24 (prepared)
431
+
432
+ - Declare the exact `remote.githubCopilotSearchRouting` dependency in the search UI child Fiber so the Web search card can render under Cordis service tracing (#137).
433
+ - Regress the actual Client apply/render callback with real Cordis traced services on both the Models footer and legacy section, including independent account activation and routing-service loss cleanup.
434
+ - Preserve all account, cancellation and paid-fallback guards. Isolated tests do not claim packaged Desktop activation or live search success.
435
+
436
+ ## 0.4.0-alpha.23 (prepared)
437
+
438
+ - Fix Model roles loading by exposing the dedicated Host `view`, `save` and `create` methods through the public Typert Remote service, retaining strict validation, CAS and session-creation ownership (#134).
439
+ - Replace Copilot-specific Auto/fixed labels and static provider suggestions with a real registered search-provider catalog shared by the primary and final-fallback selectors (#135).
440
+ - Follow the initiating Chat provider in Auto mode, keep an explicitly selected primary independent of Chat, and attempt at most one distinct final fallback. Preserve legacy settings without automatic writes and keep model choices provider-owned.
441
+ - Capture registration and account continuity before asynchronous work; silently revoked Copilot proof, cancellation, unload or registration replacement cannot authorize a paid fallback.
442
+ - Add real Host Gateway, registration-lifecycle, provider-routing and UI regressions plus an isolated built-component browser fixture. Preserve alpha.21 shared-peer ownership and alpha.22 React Client external declarations; no Core or live-profile changes are included.
443
+
444
+ ## 0.4.0-alpha.22 (prepared)
445
+
446
+ - Declare React as the DSH Client ModuleLoader external used by the actual built Client bundle, while removing it from the strict Desktop Node peer graph and retaining it only for development.
447
+ - Extend packed-manifest validation from Host shared-package intersection to every declared required peer and Client external, preserving required authorization/schemastery Host peers.
448
+ - Regress the actual packaged Desktop 0.1.5 startup failure `requires missing react@^18.2.0` without bundling a second React instance or weakening `autoInstallPeers: false` graph validation.
449
+
450
+ ## 0.4.0-alpha.21 (prepared)
451
+
452
+ - Move Desktop-owned `@deepseek-ai/dsh-authorization` and `@deepseek-ai/schemastery` from private runtime dependencies to required compatible peers while retaining development copies for standalone build and test verification.
453
+ - Audit every declared dependency and peer against the hash-pinned actual Desktop 0.1.5 runtime descriptor and generated Desktop 0.1.6 package-set input.
454
+ - Make packed-tarball verification reject bundled, optional, incompatible, or newly unaudited shared-package ownership without weakening Desktop validation or changing Copilot lifecycle, image-offload, Models, or authorization behavior.
455
+
456
+ ## 0.4.0-alpha.20 (prepared)
457
+
458
+ - Add opt-in, bilingual **Model roles** settings with account-discovered planning and execution models, revision-checked save, and an explicit new-session entry.
459
+ - Capture the two roles only for dedicated new sessions; preserve existing conversations, global defaults, OAuth ownership and ordinary Subagent settings.
460
+ - Delegate implementation through native continuable executors with an exact route, scoped workflow controls, durable policy replay and explicit model-unavailable errors rather than fallback.
461
+ - Preserve create-request identity across uncertain results, and distinguish a confirmed not-created result from failed recovery of an existing session.
462
+ - Add real React DOM, actual Client Gateway, and Core Session/projection/tool primitive regressions, plus an isolated built-component browser fixture. These tests do not imply a live model call or production Desktop activation.
463
+
464
+ ## 0.4.0-alpha.19
465
+
466
+ - Adapt the plugin compatibility contract to DSH `0.1.6-alpha.1` at
467
+ `0a15e36e7f82b6ed45af6fa9759f29b40dcd965d`.
468
+ - Exercise awaited, serialized `agent/created` initialization before reading
469
+ live Session state in the unchanged tagged-source fixture.
470
+ - Verify the 0.1.6 Session projection/history boundary, MCP v2 resource
471
+ pagination, PTC runtime and `workflow-ptc` names, cancellable Sandbox/Shell
472
+ preparation, optional-plugin startup policy, attachment cache separation,
473
+ and Team task pagination without taking ownership of those Core services.
474
+ - Follow the 0.1.6 image-budget contract: preserve `IMAGE_OFFLOAD_REQUIRED`,
475
+ apply the Core durable `image/offload` projection, and prove the retried
476
+ Copilot request uses the mapped read-only normalized path without image bytes.
477
+ - Preserve provider-scoped Copilot tool-schema filtering and the existing
478
+ immutable GitHub Release plus npm OIDC distribution design. This version is
479
+ prepared for a Draft compatibility PR only; it is not published by this change.
@@ -0,0 +1,46 @@
1
+ # Contributing
2
+
3
+ ## Scope first
4
+
5
+ Read [AGENTS.md](./AGENTS.md) before changing code. This repository is a narrow companion to DSH's built-in `@deepseek-ai/dsh-llm-pi-ai`; do not add a second general Copilot adapter, catalog, credential store, external gateway, or sandbox owner. Exact temporary model compatibility remains constrained by the documented account/ownership gates.
6
+
7
+ Use a tracking issue to describe the problem and owning seam. Security-sensitive reports belong in [private vulnerability reporting](./SECURITY.md), not a public issue. Use native DSH tools, local Git and the GitHub REST API; no external orchestrator or special VCS tool is required.
8
+
9
+ ## Plugin-only delivery
10
+
11
+ Follow the **plugin-only** boundary in [AGENTS.md](./AGENTS.md#plugin-only-implementation-boundary). Fixes must live in this plugin and use existing published public APIs. Do not modify Core source, patch installed Core or `node_modules`, monkey-patch Core internals/shared catalogs, or prepare Core PRs/releases as part of this work. A new Core export or patch being merged is not an acceptable plugin delivery prerequisite.
12
+
13
+ Core inspection is read-only. Isolated tests may use unchanged pinned Core artifacts and owned temporary fixtures, never edits to tracked Core implementation or a live deployment. If an API is insufficient, document the limitation and a plugin-local alternative; do not silently move the task into Core. A separate, explicit human Core task is required for any future Core work—generic compatibility or optimization requests do not authorize it.
14
+
15
+ ## Change workflow
16
+
17
+ 1. Determine `pwd`, Git status, remote and actual default branch. Preserve other people's changes. Start a feature branch such as `feature/issue-73-agent-readiness`; branch names need not contain an owner name.
18
+ 2. Read `node scripts/agent.mjs describe --json` and run `node scripts/agent.mjs doctor --json`. These are read-only repository diagnostics, not live health checks. Plans from `node scripts/agent.mjs plan <task> --json` contain unexecuted argv lists.
19
+ 3. Use Node 24 LTS for development (runtime floor: 22.19.0), the pinned pnpm version, and frozen install inside the checkout/worktree.
20
+ 4. Add focused regression tests before changing capability claims. Keep READMEs, deployment evidence and the Agent task map aligned.
21
+ 5. State expected outcomes and scope; run:
22
+
23
+ ```sh
24
+ pnpm install --frozen-lockfile
25
+ pnpm verify
26
+ pnpm pack --pack-destination artifacts
27
+ ```
28
+
29
+ Validate the resulting exact path with `pnpm verify:tarball -- artifacts/dsh-github-copilot-<package-version>.tgz`. The release task plan supplies that path automatically. Never run broad cleanup on an existing artifacts directory just to package again.
30
+ 6. Inspect the complete diff, compare expected versus actual results, and report what was NOT checked. Source markers, stubbed unit tests, built module imports, upstream config fixtures, authenticated API calls, installation, and loaded UI are different evidence layers.
31
+ 7. Commit with the issue reference and accurate attribution. For work assisted by DSH use `Assisted-by: DeepSeek Harness (DSH)`. Other agents must name the actual tool; do not infer authorship from the model vendor or copy a bot identity. Add `Co-authored-by` only for real verified collaborators. No fabricated email, author replacement or published-history rewrite.
32
+ 8. Push the feature branch and open a PR containing `Fixes #<issue>`, verification, risks and rollback. For user-requested delivery, merge without an additional approval prompt after required review/CI and branch protection; respect explicit code-only, review-only or do-not-merge restrictions.
33
+
34
+ Generated `lib/` output, archives, `.env`, tokens, local credentials, machine state and production screenshots with real authorization codes must not be committed. Use synthetic fixtures for screenshots and report staged UI honestly.
35
+
36
+ ## Release changes
37
+
38
+ GitHub Releases and npm are the default distribution channels for every new version, superseding the Release-only decision in #43 at the user's request. The public package uses the exact same verified tarball in both channels. No automatic npm opt-out is allowed. See [npm distribution](./docs/npm-distribution.md) for initial maintainer bootstrap, OIDC, retry recovery and organizational access requirements.
39
+
40
+ Important user-requested features, behavior fixes, compatibility fixes, and security or stability fixes include release follow-through by default after a reviewed merge and green required CI. Do not stop at the merged PR or ask for a second release confirmation. User instructions limiting delivery to code/review or forbidding publication take precedence; documentation-only, internal-only and unrelated releases are not implicitly authorized. Follow the [important-update delivery rule](./AGENTS.md#important-update-release-delivery).
41
+
42
+ Prepare version alignment in the implementation PR. Important runtime or delivery changes fail the `release-ready` PR check unless `package.json` is strictly newer than the base and `deployment-baseline.json` plus both README install URLs agree. Retain the current prerelease channel unless promotion is requested. Docs/tests-only changes may keep the current version. Never move or reuse a tag. Local installation, sign-out and worktree checkout still need explicit approval; restarting active Sessions requires acknowledgement of the interruption.
43
+
44
+ After the release-ready PR merges, the successful main CI run calls the reusable Release workflow on that exact SHA. It creates or reconciles the annotated version tag and a draft-first Release, rebuilds and verifies the package, writes and verifies `SHA256SUMS`, checks remote asset digests, and only then publishes an immutable Release. Identical reruns reconcile safely; conflicting tags, releases or bytes fail closed. A tag created by `GITHUB_TOKEN` does not trigger a second workflow, so publication occurs in the gated main workflow rather than depending on recursive tag events. Do not hand-upload a replacement or bypass failed CI.
45
+
46
+ Release-bearing work is complete only after reporting the published Release URL and verified version/commit/assets/checksum, or clearly identifying a concrete publication blocker and pending step. A merged PR or a local tarball alone is not release delivery. After publication, independently verify the Release is not a draft, its prerelease flag, exact tag/commit, asset names and downloaded SHA-256. An install command must include the target `--profile`; never overwrite a shipped preset. Compare the installed package to the downloaded asset, then separately validate activation after restart. HTTP 200 alone does not prove the new plugin loaded.
package/LICENSE ADDED
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2025 yms
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.