@yugoo-design/companion 0.1.0-beta.12 → 0.1.0-beta.13

package/README.md

@@ -1,6 +1,6 @@
 # Yugoo Design Companion — beta
 
-Package: `@yugoo-design/companion@0.1.0-beta.12`.
+Package: `@yugoo-design/companion@0.1.0-beta.13`.
 
 Yugoo Design (语构设计系统) is an interactive design plugin for native Agent Hosts.
 It combines domain methods, Design Projects, versioned Design Files and shared
@@ -21,7 +21,7 @@ selected beta to the exact version and integrity before acquisition; keep the
 previous installed version available during updates. Do not substitute `latest`.
 
 ```sh
-ONNXRUNTIME_NODE_INSTALL_CUDA=skip npm install --ignore-scripts=false --save-exact @yugoo-design/companion@0.1.0-beta.12 --registry=https://registry.npmjs.org
+ONNXRUNTIME_NODE_INSTALL_CUDA=skip npm install --ignore-scripts=false --save-exact @yugoo-design/companion@0.1.0-beta.13 --registry=https://registry.npmjs.org
 npx --no-install design-system --help
 ```
 
@@ -38,7 +38,7 @@ allowBuilds:
 ```
 
 ```sh
-ONNXRUNTIME_NODE_INSTALL_CUDA=skip pnpm add --save-exact @yugoo-design/companion@0.1.0-beta.12 --registry=https://registry.npmjs.org
+ONNXRUNTIME_NODE_INSTALL_CUDA=skip pnpm add --save-exact @yugoo-design/companion@0.1.0-beta.13 --registry=https://registry.npmjs.org
 pnpm exec design-system --help
 ```
 
@@ -86,7 +86,7 @@ process value:
 $previousOrtCuda = [Environment]::GetEnvironmentVariable('ONNXRUNTIME_NODE_INSTALL_CUDA', 'Process')
 try {
   $env:ONNXRUNTIME_NODE_INSTALL_CUDA = 'skip'
-  npm install --ignore-scripts=false --save-exact @yugoo-design/companion@0.1.0-beta.12 --registry=https://registry.npmjs.org
+  npm install --ignore-scripts=false --save-exact @yugoo-design/companion@0.1.0-beta.13 --registry=https://registry.npmjs.org
 } finally {
   [Environment]::SetEnvironmentVariable('ONNXRUNTIME_NODE_INSTALL_CUDA', $previousOrtCuda, 'Process')
 }
@@ -108,14 +108,18 @@ to an already published package. Users do not install DSH and Companion separate
 | WorkBuddy | `install --host workbuddy`: prepare the stable marketplace and plugin; import or update and enable through the host. |
 | DSH | `install --host dsh`: run the existing plan/apply transaction with the bound aggregate, official CLI/Loader and an explicitly selected existing profile. |
 
-For a fresh WorkBuddy installation, use native suite `0.3.25` with
-`--suite-version 0.3.25`, the verified Node executable and a persistent external
+For a fresh WorkBuddy installation, use native suite `0.3.26` with
+`--suite-version 0.3.26`, the verified Node executable and a persistent external
 preparation destination. Import the returned `nativeImport.market` directory and
 enable its `yugoo-design` plugin. The fresh market is also named `yugoo-design`;
 no previous test market or preparation receipt is required. An owned update with
 `--previous-receipt` retains its original market name and directory, including
 historical test markets, rather than migrating a registered installation.
 
+For QwenWork, use native suite `0.1.2` with `--suite-version 0.1.2`.
+Its suite version advances independently of the WorkBuddy suite and Companion
+npm version; desktop preparation writes that explicit version into the import.
+
 Supply the required host inputs shown by `design-system --help`. Without a host,
 the entry returns `host-selection-required`; missing inputs return
 `inputs-required`, both with exit code 2 and no installation attempted. The

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@yugoo-design/companion",
-  "version": "0.1.0-beta.12",
+  "version": "0.1.0-beta.13",
   "description": "Yugoo Design Companion beta: interactive Design Projects and Panels in supported native Agent Hosts.",
   "license": "Apache-2.0",
   "type": "module",

package/payload/desktop-preparation/hosts/workbuddy/design-SKILL.md

@@ -151,11 +151,24 @@ Only `backgroundReviewReadiness: ready` and `backgroundReviews: true` permit
 dispatch. WorkBuddy 5.8.3's public `Agent(run_in_background=true)` can fall back
 to synchronous execution when native background capability is disabled. Do not
 probe it with a real review prompt. Unverified readiness keeps requests pending;
-use the returned content-free `backgroundReviewProtocol.bootstrap` only as an
-explicit readiness check, then read status again. Do not change Host settings or
-repeat an uncertain probe. Proof must be newer than this exact Companion connection
+use the returned content-free `backgroundReviewProtocol.bootstrap` only when
+`canAttemptReadiness` is true, as one explicit readiness check, then read status
+again. When false, inspect the existing result and reported reason; do not launch
+another probe. A missing native transport observation does not prove synchronous
+execution; a verified synchronous child result cannot prove background readiness.
+Do not change Host settings or repeat an uncertain probe. Proof must be newer than this exact Companion connection
 opening; a historical descriptor from the same session is insufficient.
 
+If setup is missing, WorkBuddy 5.8.3's Chinese desktop exposes avatar → 设置 →
+智能体 → 智能体团队管理 → 开启. Ask the user to inspect it and decide whether to
+enable native multi-agent collaboration; never change Host settings or an expert
+manifest yourself. Do not infer the current value from the startup default or a
+synchronous fallback. The switch alone is not a verified background launch, and
+this guide does not require a full Host restart. After normal plugin activation
+or user-controlled setup, use the exact new conversation/connection and its
+actual readiness contract. Foreground design, KAG and main U/S/T remain available
+independently of GRA readiness.
+
 When ready, honor the project GRA policy (`off` forbids automatic review) and reuse
 the exact `dispatch:true` prepared IDs. Follow the installed native execution
 reference: launch independent read-only native background reviewers, bind actual

package/payload/desktop-preparation/materials.json

@@ -5,9 +5,9 @@
     "LICENSE": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30",
     "NOTICE": "6b4eec391144fcbd96ec7b45a28764ab1932722a96325248377c3e28aac2c62d",
     "THIRD_PARTY_NOTICES.md": "bfcab921c6c15ccc47f9b1a6cc5941675c95962de5ea5b0edc50fbfa2aef136a",
-    "hosts/workbuddy/design-SKILL.md": "5cc44054186fccb6bba66e237367cec3774ea59f64ae93ea4acf5648355f22b0",
-    "plugin/design-system/skills/design-panel/native-execution.md": "7a2ab8b2aff916115a8c767c3c1e8156dc71a58c154366a0e8ba5c80256aeda3",
-    "qwenwork-package.mjs": "ae340c245feb2beab0de9e02c09b37d9924fa0d6fe9f3dd1c329242e4a350c11",
+    "hosts/workbuddy/design-SKILL.md": "6f49c7d62a658c86a679dfeff486c9edf11c83095c17d6290ee03ccc933020a6",
+    "plugin/design-system/skills/design-panel/native-execution.md": "daa9696fef481bd9da827aba963ee4296eb288d8f72c56494fcc670e29f0e86c",
+    "qwenwork-package.mjs": "b43c2e9a7c3476497a3c81e143d36059cf0ce9509536aa882635cb370f2413db",
     "qwenwork-template/.claude-plugin/plugin.json": "7beaf01908b62b4ad6c0a371f09b45cb3236f622910c3769f8b3393e4cacb354",
     "qwenwork-template/.mcp.json": "931fdcd8efa8e0cfe1dbdda225c523d62dbc4bacb5a33d01eebe84933b712353",
     "qwenwork-template/LICENSE": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30",
@@ -23,7 +23,7 @@
     "qwenwork-template/scripts/windows-private-storage.mjs": "27c3eb052278f0cb2850326e1d0864cfd2f3f6d2b1975ec7c24535ef6251ac94",
     "qwenwork-template/skills/design-component/SKILL.md": "85a1f2a016174787226eaa9e9e6675e86057212d322a65014edd71f3e4861b34",
     "qwenwork-template/skills/yugoo/SKILL.md": "3aedb6739f204c9931cf1d406e5a71d59fc3424875746ee5737e68e5a58236bc",
-    "qwenwork-template/skills/yugoo/native-execution.md": "9d300490b3493ff9f4826414cfa832244acc29b2205b825d3ccaff8665049ddd",
+    "qwenwork-template/skills/yugoo/native-execution.md": "b4cbc2e7621ac8c9a89ad58f05f9a46b3b3ae241b290e7386728cb1a4aa2a8b2",
     "src/browser-access.mjs": "4b64284b9bfaaa32c8502ab00febdda4f8a972f4b41e822d4e708b27e7e3efdf",
     "src/native-hook-transport.mjs": "a5878335c40bc930d28d34e079de09f7628dc0cbcc828a1309826c1829640a83",
     "src/qwenwork-connection.mjs": "850b95ea4f57ccf4e1cb6463f98625b941d74f4df76bc24f05dbdf15b8597b3a",
@@ -37,6 +37,6 @@
     "src/workbuddy-relay-probe.mjs": "c5ea46abadbf9dfe265eb40eb81ed08a4f6596dd5f673fd248843e7775160cc4",
     "src/workbuddy-relay-service.mjs": "44d60d6956250417507fb76c71c3030a01bc38e565e4572a761b46f255f8e669",
     "src/workbuddy-ui-probes.mjs": "a91d4651248125769254755bdf66a44bbc9e5a793dad0a137c35acbb10a89a88",
-    "workbuddy-package.mjs": "cc3ca0be5cc120f107119bf82897a83d555f71338347c7908568dba562111ec9"
+    "workbuddy-package.mjs": "9ec792fd0c56ccfe76a3cbf6215b38999677b7ce761ab2f9a95698684d6995ca"
   }
 }

package/payload/desktop-preparation/plugin/design-system/skills/design-panel/native-execution.md

@@ -2,7 +2,8 @@
 
 The native Host owns spawning, concurrency, messages, cancellation and permissions. The
 companion owns Design Files, Panels and the existing validation/Git/journal path.
-Do not invoke a second `codex exec`, `resume`, or app-server to simulate children.
+Do not invoke a second model client, CLI process or app-server to simulate children;
+for example, a second `codex exec`/`resume` is not native Codex delegation.
 If native delegation is unavailable, report the limitation rather than claim
 fanout succeeded. The Producer may perform the scoped artifact work itself.
 
@@ -190,7 +191,7 @@ settles its Bundle. Do not hide failure by substituting another file.
 
 ## Mapping existing Bundle tool terminology
 
-| Bundle operation | Native Codex operation |
+| Bundle operation | Native Host operation |
 | --- | --- |
 | `design_delegate` | Native child spawn, followed by exact run/target registration |
 | `read_skill_resource` | Read the referenced file under returned Bundle resources |
@@ -269,6 +270,22 @@ structured `agent-output` receipt supplies native status. Externalized/missing
 status or newer unbound child activity remains unknown. Do not follow paths in
 tool text, busy-poll, wait for review completion or block the main user task.
 
+QwenWork has a separate native background contract. Read its current
+`hostCapabilities.backgroundReviewReadiness` and `backgroundReviewProtocol`;
+only `ready` permits real review dispatch. An explicitly requested content-free
+bootstrap may establish readiness after this exact connection opened, but never
+repeat it automatically or change Host settings. Its structured parent
+`tool_result` must contain `isAsync: true` and `status: "async_launched"`, matched
+to that exact native Agent call and the Host's task descriptor. A textual Agent
+ID, synchronous fallback or older task is insufficient. Each actual review must
+independently satisfy this proof when bound; bootstrap success is not a review
+launch. Bind the actual native `agentId` as `runId`, plus its exact `taskPath` and
+child `transcriptPath`. The QwenWork adapter observes native task status and
+child evidence; do not copy WorkBuddy's version-1 descriptor or `TaskOutput`
+contract, infer completion from the Agent RPC, busy-poll or wait in the foreground.
+Unknown/unreadable native evidence remains unavailable. The installed-version
+contract and limitations live in the QwenWork integration entry.
+
 When `design_status` reports `reviewDispatchOwner: "companion"`, or preparation
 returns `dispatchOwner: "companion"`, the service owns background review execution.
 Do not prepare, spawn, bind, follow up, or wait for reviewer children. Continue the
@@ -328,7 +345,12 @@ the parent need not read or rewrite them. Do not dump Design Files into the pare
 
 Spawn independent native **background** review children from the main Agent.
 Give each only its review ID, connection path and instructions to read its request
-and submit its own structured result. Launch independent children together when
+and submit its own structured result. Include that exact review ID literally in
+the native Agent prompt. WorkBuddy and QwenWork verify both the current launch
+boundary and this request assignment on first binding; an old task or READY
+probe cannot become a new Reviewer, even when it belongs to the same session.
+Exact already-bound historical evidence remains readable after reopening.
+Launch independent children together when
 the host supports it. Use one `bindReviewRuns` call with
 `{bindings:[{reviewId,nativeRun},...]}` for the exact native launch results;
 `nativeRun` can contain just the actual `runId` (or, for `host=codex` only, an exact launch reference
@@ -363,7 +385,7 @@ The reviewers have no authority to edit Design Files or automatically accept
 advice. Use `submitReview` for structured result publication, preserving the
 exact review ID/native child binding. Errors return diagnostics for correction.
 Do not force stale results onto newer project evidence. UI reads and polling
-must never launch reviews. Native permission/cancellation remains in Codex.
+must never launch reviews. Native permission/cancellation remains with the Host.
 
 GRA and U/S/T are distinct: U/S/T updates assessment; GRA contributes suggestions
 to the shared advisory inbox. Only a real user acceptance dispatches a scoped

package/payload/desktop-preparation/qwenwork-package.mjs

@@ -2,7 +2,7 @@
 import { readFile, writeFile } from 'node:fs/promises'
 import { join, win32 } from 'node:path'
 export function qwenWorkNativeExecutionReference(shared) {
-  return '# QwenWork native execution\n\nRead SKILL.md for exact native task descriptor and user authorization requirements.\n\n' + shared.replaceAll('Codex', 'QwenWork').replace('`codex exec`', 'a native QwenWork agent')
+  return '# QwenWork native execution\n\nRead SKILL.md for exact native task descriptor and user authorization requirements. The shared reference below retains explicit host qualifiers: Codex-only launch references are not QwenWork parameters. Use the QwenWork background contract, never another Host\'s receipt format.\n\n' + shared
 }
 export function validateQwenWorkRuntime(value) {
   if (!value || !['executable', 'worker', 'nodeExecutable'].every(k => typeof value[k] === 'string' && win32.isAbsolute(value[k]) && !/[\r\n\0]/.test(value[k]))) throw Error('QwenWork requires explicit Windows executable, worker and Node paths.')

package/payload/desktop-preparation/qwenwork-template/skills/yugoo/native-execution.md

@@ -1,12 +1,13 @@
 # QwenWork native execution
 
-Read SKILL.md for exact native task descriptor and user authorization requirements.
+Read SKILL.md for exact native task descriptor and user authorization requirements. The shared reference below retains explicit host qualifiers: Codex-only launch references are not QwenWork parameters. Use the QwenWork background contract, never another Host's receipt format.
 
 # Native Design execution
 
 The native Host owns spawning, concurrency, messages, cancellation and permissions. The
 companion owns Design Files, Panels and the existing validation/Git/journal path.
-Do not invoke a second a native QwenWork agent, `resume`, or app-server to simulate children.
+Do not invoke a second model client, CLI process or app-server to simulate children;
+for example, a second `codex exec`/`resume` is not native Codex delegation.
 If native delegation is unavailable, report the limitation rather than claim
 fanout succeeded. The Producer may perform the scoped artifact work itself.
 
@@ -194,7 +195,7 @@ settles its Bundle. Do not hide failure by substituting another file.
 
 ## Mapping existing Bundle tool terminology
 
-| Bundle operation | Native QwenWork operation |
+| Bundle operation | Native Host operation |
 | --- | --- |
 | `design_delegate` | Native child spawn, followed by exact run/target registration |
 | `read_skill_resource` | Read the referenced file under returned Bundle resources |
@@ -273,6 +274,22 @@ structured `agent-output` receipt supplies native status. Externalized/missing
 status or newer unbound child activity remains unknown. Do not follow paths in
 tool text, busy-poll, wait for review completion or block the main user task.
 
+QwenWork has a separate native background contract. Read its current
+`hostCapabilities.backgroundReviewReadiness` and `backgroundReviewProtocol`;
+only `ready` permits real review dispatch. An explicitly requested content-free
+bootstrap may establish readiness after this exact connection opened, but never
+repeat it automatically or change Host settings. Its structured parent
+`tool_result` must contain `isAsync: true` and `status: "async_launched"`, matched
+to that exact native Agent call and the Host's task descriptor. A textual Agent
+ID, synchronous fallback or older task is insufficient. Each actual review must
+independently satisfy this proof when bound; bootstrap success is not a review
+launch. Bind the actual native `agentId` as `runId`, plus its exact `taskPath` and
+child `transcriptPath`. The QwenWork adapter observes native task status and
+child evidence; do not copy WorkBuddy's version-1 descriptor or `TaskOutput`
+contract, infer completion from the Agent RPC, busy-poll or wait in the foreground.
+Unknown/unreadable native evidence remains unavailable. The installed-version
+contract and limitations live in the QwenWork integration entry.
+
 When `design_status` reports `reviewDispatchOwner: "companion"`, or preparation
 returns `dispatchOwner: "companion"`, the service owns background review execution.
 Do not prepare, spawn, bind, follow up, or wait for reviewer children. Continue the
@@ -332,7 +349,12 @@ the parent need not read or rewrite them. Do not dump Design Files into the pare
 
 Spawn independent native **background** review children from the main Agent.
 Give each only its review ID, connection path and instructions to read its request
-and submit its own structured result. Launch independent children together when
+and submit its own structured result. Include that exact review ID literally in
+the native Agent prompt. WorkBuddy and QwenWork verify both the current launch
+boundary and this request assignment on first binding; an old task or READY
+probe cannot become a new Reviewer, even when it belongs to the same session.
+Exact already-bound historical evidence remains readable after reopening.
+Launch independent children together when
 the host supports it. Use one `bindReviewRuns` call with
 `{bindings:[{reviewId,nativeRun},...]}` for the exact native launch results;
 `nativeRun` can contain just the actual `runId` (or, for `host=codex` only, an exact launch reference
@@ -367,7 +389,7 @@ The reviewers have no authority to edit Design Files or automatically accept
 advice. Use `submitReview` for structured result publication, preserving the
 exact review ID/native child binding. Errors return diagnostics for correction.
 Do not force stale results onto newer project evidence. UI reads and polling
-must never launch reviews. Native permission/cancellation remains in QwenWork.
+must never launch reviews. Native permission/cancellation remains with the Host.
 
 GRA and U/S/T are distinct: U/S/T updates assessment; GRA contributes suggestions
 to the shared advisory inbox. Only a real user acceptance dispatches a scoped

package/payload/desktop-preparation/workbuddy-package.mjs

@@ -3,7 +3,7 @@ import { mkdir, writeFile, copyFile } from 'node:fs/promises'
 import { join, win32 } from 'node:path'
 import { resolveBrowserAccess } from './src/browser-access.mjs'
 
-export const WORKBUDDY_TEST_VERSION = '0.3.25'
+export const WORKBUDDY_TEST_VERSION = '0.3.26'
 export const WORKBUDDY_TEST_MARKET = 'yugoo-design'
 export const WORKBUDDY_TEST_PLUGIN = 'yugoo-design'
 

package/payload/overlays/claude-code/.claude-plugin/plugin.json

@@ -1,5 +1,5 @@
 {
   "name": "yugoo-design",
-  "version": "0.1.0-beta.12",
+  "version": "0.1.0-beta.13",
   "description": "Yugoo Design: interactive Design Panels with Claude Code native agents and background design reviews."
 }

package/payload/overlays/claude-code/skills/yugoo/native-execution.md

@@ -2,7 +2,8 @@
 
 The native Host owns spawning, concurrency, messages, cancellation and permissions. The
 companion owns Design Files, Panels and the existing validation/Git/journal path.
-Do not invoke a second `codex exec`, `resume`, or app-server to simulate children.
+Do not invoke a second model client, CLI process or app-server to simulate children;
+for example, a second `codex exec`/`resume` is not native Codex delegation.
 If native delegation is unavailable, report the limitation rather than claim
 fanout succeeded. The Producer may perform the scoped artifact work itself.
 
@@ -190,7 +191,7 @@ settles its Bundle. Do not hide failure by substituting another file.
 
 ## Mapping existing Bundle tool terminology
 
-| Bundle operation | Native Codex operation |
+| Bundle operation | Native Host operation |
 | --- | --- |
 | `design_delegate` | Native child spawn, followed by exact run/target registration |
 | `read_skill_resource` | Read the referenced file under returned Bundle resources |
@@ -269,6 +270,22 @@ structured `agent-output` receipt supplies native status. Externalized/missing
 status or newer unbound child activity remains unknown. Do not follow paths in
 tool text, busy-poll, wait for review completion or block the main user task.
 
+QwenWork has a separate native background contract. Read its current
+`hostCapabilities.backgroundReviewReadiness` and `backgroundReviewProtocol`;
+only `ready` permits real review dispatch. An explicitly requested content-free
+bootstrap may establish readiness after this exact connection opened, but never
+repeat it automatically or change Host settings. Its structured parent
+`tool_result` must contain `isAsync: true` and `status: "async_launched"`, matched
+to that exact native Agent call and the Host's task descriptor. A textual Agent
+ID, synchronous fallback or older task is insufficient. Each actual review must
+independently satisfy this proof when bound; bootstrap success is not a review
+launch. Bind the actual native `agentId` as `runId`, plus its exact `taskPath` and
+child `transcriptPath`. The QwenWork adapter observes native task status and
+child evidence; do not copy WorkBuddy's version-1 descriptor or `TaskOutput`
+contract, infer completion from the Agent RPC, busy-poll or wait in the foreground.
+Unknown/unreadable native evidence remains unavailable. The installed-version
+contract and limitations live in the QwenWork integration entry.
+
 When `design_status` reports `reviewDispatchOwner: "companion"`, or preparation
 returns `dispatchOwner: "companion"`, the service owns background review execution.
 Do not prepare, spawn, bind, follow up, or wait for reviewer children. Continue the
@@ -328,7 +345,12 @@ the parent need not read or rewrite them. Do not dump Design Files into the pare
 
 Spawn independent native **background** review children from the main Agent.
 Give each only its review ID, connection path and instructions to read its request
-and submit its own structured result. Launch independent children together when
+and submit its own structured result. Include that exact review ID literally in
+the native Agent prompt. WorkBuddy and QwenWork verify both the current launch
+boundary and this request assignment on first binding; an old task or READY
+probe cannot become a new Reviewer, even when it belongs to the same session.
+Exact already-bound historical evidence remains readable after reopening.
+Launch independent children together when
 the host supports it. Use one `bindReviewRuns` call with
 `{bindings:[{reviewId,nativeRun},...]}` for the exact native launch results;
 `nativeRun` can contain just the actual `runId` (or, for `host=codex` only, an exact launch reference
@@ -363,7 +385,7 @@ The reviewers have no authority to edit Design Files or automatically accept
 advice. Use `submitReview` for structured result publication, preserving the
 exact review ID/native child binding. Errors return diagnostics for correction.
 Do not force stale results onto newer project evidence. UI reads and polling
-must never launch reviews. Native permission/cancellation remains in Codex.
+must never launch reviews. Native permission/cancellation remains with the Host.
 
 GRA and U/S/T are distinct: U/S/T updates assessment; GRA contributes suggestions
 to the shared advisory inbox. Only a real user acceptance dispatches a scoped

package/payload/overlays/codex/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "yugoo-design",
-  "version": "0.1.0-beta.12",
+  "version": "0.1.0-beta.13",
   "description": "Yugoo Design: local Design Project Panels for an explicitly bound Codex conversation (development preview).",
   "author": {
     "name": "yugoo-design contributors"

package/payload/overlays/codex/skills/design-panel/native-execution.md

@@ -2,7 +2,8 @@
 
 The native Host owns spawning, concurrency, messages, cancellation and permissions. The
 companion owns Design Files, Panels and the existing validation/Git/journal path.
-Do not invoke a second `codex exec`, `resume`, or app-server to simulate children.
+Do not invoke a second model client, CLI process or app-server to simulate children;
+for example, a second `codex exec`/`resume` is not native Codex delegation.
 If native delegation is unavailable, report the limitation rather than claim
 fanout succeeded. The Producer may perform the scoped artifact work itself.
 
@@ -190,7 +191,7 @@ settles its Bundle. Do not hide failure by substituting another file.
 
 ## Mapping existing Bundle tool terminology
 
-| Bundle operation | Native Codex operation |
+| Bundle operation | Native Host operation |
 | --- | --- |
 | `design_delegate` | Native child spawn, followed by exact run/target registration |
 | `read_skill_resource` | Read the referenced file under returned Bundle resources |
@@ -269,6 +270,22 @@ structured `agent-output` receipt supplies native status. Externalized/missing
 status or newer unbound child activity remains unknown. Do not follow paths in
 tool text, busy-poll, wait for review completion or block the main user task.
 
+QwenWork has a separate native background contract. Read its current
+`hostCapabilities.backgroundReviewReadiness` and `backgroundReviewProtocol`;
+only `ready` permits real review dispatch. An explicitly requested content-free
+bootstrap may establish readiness after this exact connection opened, but never
+repeat it automatically or change Host settings. Its structured parent
+`tool_result` must contain `isAsync: true` and `status: "async_launched"`, matched
+to that exact native Agent call and the Host's task descriptor. A textual Agent
+ID, synchronous fallback or older task is insufficient. Each actual review must
+independently satisfy this proof when bound; bootstrap success is not a review
+launch. Bind the actual native `agentId` as `runId`, plus its exact `taskPath` and
+child `transcriptPath`. The QwenWork adapter observes native task status and
+child evidence; do not copy WorkBuddy's version-1 descriptor or `TaskOutput`
+contract, infer completion from the Agent RPC, busy-poll or wait in the foreground.
+Unknown/unreadable native evidence remains unavailable. The installed-version
+contract and limitations live in the QwenWork integration entry.
+
 When `design_status` reports `reviewDispatchOwner: "companion"`, or preparation
 returns `dispatchOwner: "companion"`, the service owns background review execution.
 Do not prepare, spawn, bind, follow up, or wait for reviewer children. Continue the
@@ -328,7 +345,12 @@ the parent need not read or rewrite them. Do not dump Design Files into the pare
 
 Spawn independent native **background** review children from the main Agent.
 Give each only its review ID, connection path and instructions to read its request
-and submit its own structured result. Launch independent children together when
+and submit its own structured result. Include that exact review ID literally in
+the native Agent prompt. WorkBuddy and QwenWork verify both the current launch
+boundary and this request assignment on first binding; an old task or READY
+probe cannot become a new Reviewer, even when it belongs to the same session.
+Exact already-bound historical evidence remains readable after reopening.
+Launch independent children together when
 the host supports it. Use one `bindReviewRuns` call with
 `{bindings:[{reviewId,nativeRun},...]}` for the exact native launch results;
 `nativeRun` can contain just the actual `runId` (or, for `host=codex` only, an exact launch reference
@@ -363,7 +385,7 @@ The reviewers have no authority to edit Design Files or automatically accept
 advice. Use `submitReview` for structured result publication, preserving the
 exact review ID/native child binding. Errors return diagnostics for correction.
 Do not force stale results onto newer project evidence. UI reads and polling
-must never launch reviews. Native permission/cancellation remains in Codex.
+must never launch reviews. Native permission/cancellation remains with the Host.
 
 GRA and U/S/T are distinct: U/S/T updates assessment; GRA contributes suggestions
 to the shared advisory inbox. Only a real user acceptance dispatches a scoped

package/payload/overlays/qoder/.qoder-plugin/plugin.json

@@ -1,5 +1,5 @@
 {
   "name": "yugoo-design",
-  "version": "0.1.0-beta.12",
+  "version": "0.1.0-beta.13",
   "description": "Yugoo Design: interactive Design Panels with Qoder-native Producers, Workers and background reviews."
 }

package/payload/overlays/qoder/skills/yugoo/native-execution.md

@@ -6,7 +6,8 @@ Read SKILL.md native execution specifics for Qoder taskPath and agent IDs. The o
 
 The native Host owns spawning, concurrency, messages, cancellation and permissions. The
 companion owns Design Files, Panels and the existing validation/Git/journal path.
-Do not invoke a second `qodercli -p`, `resume`, or app-server to simulate children.
+Do not invoke a second model client, CLI process or app-server to simulate children;
+for example, a second `qodercli -p`/`resume` is not native Qoder CLI delegation.
 If native delegation is unavailable, report the limitation rather than claim
 fanout succeeded. The Producer may perform the scoped artifact work itself.
 
@@ -194,7 +195,7 @@ settles its Bundle. Do not hide failure by substituting another file.
 
 ## Mapping existing Bundle tool terminology
 
-| Bundle operation | Native Qoder CLI operation |
+| Bundle operation | Native Host operation |
 | --- | --- |
 | `design_delegate` | Native child spawn, followed by exact run/target registration |
 | `read_skill_resource` | Read the referenced file under returned Bundle resources |
@@ -273,6 +274,22 @@ structured `agent-output` receipt supplies native status. Externalized/missing
 status or newer unbound child activity remains unknown. Do not follow paths in
 tool text, busy-poll, wait for review completion or block the main user task.
 
+QwenWork has a separate native background contract. Read its current
+`hostCapabilities.backgroundReviewReadiness` and `backgroundReviewProtocol`;
+only `ready` permits real review dispatch. An explicitly requested content-free
+bootstrap may establish readiness after this exact connection opened, but never
+repeat it automatically or change Host settings. Its structured parent
+`tool_result` must contain `isAsync: true` and `status: "async_launched"`, matched
+to that exact native Agent call and the Host's task descriptor. A textual Agent
+ID, synchronous fallback or older task is insufficient. Each actual review must
+independently satisfy this proof when bound; bootstrap success is not a review
+launch. Bind the actual native `agentId` as `runId`, plus its exact `taskPath` and
+child `transcriptPath`. The QwenWork adapter observes native task status and
+child evidence; do not copy WorkBuddy's version-1 descriptor or `TaskOutput`
+contract, infer completion from the Agent RPC, busy-poll or wait in the foreground.
+Unknown/unreadable native evidence remains unavailable. The installed-version
+contract and limitations live in the QwenWork integration entry.
+
 When `design_status` reports `reviewDispatchOwner: "companion"`, or preparation
 returns `dispatchOwner: "companion"`, the service owns background review execution.
 Do not prepare, spawn, bind, follow up, or wait for reviewer children. Continue the
@@ -332,7 +349,12 @@ the parent need not read or rewrite them. Do not dump Design Files into the pare
 
 Spawn independent native **background** review children from the main Agent.
 Give each only its review ID, connection path and instructions to read its request
-and submit its own structured result. Launch independent children together when
+and submit its own structured result. Include that exact review ID literally in
+the native Agent prompt. WorkBuddy and QwenWork verify both the current launch
+boundary and this request assignment on first binding; an old task or READY
+probe cannot become a new Reviewer, even when it belongs to the same session.
+Exact already-bound historical evidence remains readable after reopening.
+Launch independent children together when
 the host supports it. Use one `bindReviewRuns` call with
 `{bindings:[{reviewId,nativeRun},...]}` for the exact native launch results;
 `nativeRun` can contain just the actual `runId` (or, for `host=codex` only, an exact launch reference
@@ -367,7 +389,7 @@ The reviewers have no authority to edit Design Files or automatically accept
 advice. Use `submitReview` for structured result publication, preserving the
 exact review ID/native child binding. Errors return diagnostics for correction.
 Do not force stale results onto newer project evidence. UI reads and polling
-must never launch reviews. Native permission/cancellation remains in Qoder CLI.
+must never launch reviews. Native permission/cancellation remains with the Host.
 
 GRA and U/S/T are distinct: U/S/T updates assessment; GRA contributes suggestions
 to the shared advisory inbox. Only a real user acceptance dispatches a scoped

package/payload/shared/scripts/{content-archive-ChKG0eA3.mjs → content-archive-CSdtYjTT.mjs}

@@ -1,5 +1,5 @@
-import { Dt as DesignKitVersion, Et as DesignBundleVersion, at as loadDesignBundle } from "./core-entry-Cfvnjfig.mjs";
-import { t as loadStudioKit } from "./load-C46tLUT4.mjs";
+import { Dt as DesignKitVersion, Et as DesignBundleVersion, at as loadDesignBundle } from "./core-entry-CCOXUWXc.mjs";
+import { t as loadStudioKit } from "./load-B1XGiCBF.mjs";
 import { extname, isAbsolute, relative, resolve, sep } from "node:path";
 import { createHash } from "node:crypto";
 //#region packages/studio/src/content-archive.ts

package/payload/shared/scripts/design.mjs

@@ -1,11 +1,11 @@
 import { r as __toESM, t as __commonJSMin } from "./rolldown-runtime-CMFfr-1z.mjs";
-import { a as createClaudeTranscript, c as assertCompanionKitContributions, d as requireThreadId, f as notificationErrorResponse, i as resolveCodexMainTranscript, l as createCompanionContentLibrary, n as companionKnowledgeRuntimeUrl, o as companionFs, p as unknownDeliveryError, r as runtime, s as openCompanionProject, t as startCompanion, u as createQwenWorkTranscript } from "./server-CxKbY4vh.mjs";
-import { R as validateSkillDesign, X as runDesignProjectCommand, Y as parseDesignCommand, Z as DesignProjectRegistry, gt as PANEL_COMPONENT_CATALOG, ht as GENERATED_COMPONENT_KINDS, mt as COMPONENT_KINDS } from "./core-entry-Cfvnjfig.mjs";
+import { a as createClaudeTranscript, c as assertCompanionKitContributions, d as requireThreadId, f as notificationErrorResponse, i as resolveCodexMainTranscript, l as createCompanionContentLibrary, n as companionKnowledgeRuntimeUrl, o as companionFs, p as unknownDeliveryError, r as runtime, s as openCompanionProject, t as startCompanion, u as createQwenWorkTranscript } from "./server-DZXeZst5.mjs";
+import { R as validateSkillDesign, X as runDesignProjectCommand, Y as parseDesignCommand, Z as DesignProjectRegistry, gt as PANEL_COMPONENT_CATALOG, ht as GENERATED_COMPONENT_KINDS, mt as COMPONENT_KINDS } from "./core-entry-CCOXUWXc.mjs";
 import { n as DESIGN_GIT_READ_DESCRIPTION, r as designGitReadArguments, t as DESIGN_GIT_READ_ACTIONS } from "./design-git-read-BUV0s4TT.mjs";
 import { dirname, isAbsolute, join, resolve } from "node:path";
 import { mkdir, readFile, rename, rmdir, writeFile } from "node:fs/promises";
-import { parseArgs } from "node:util";
 import { createHash, randomUUID } from "node:crypto";
+import { parseArgs } from "node:util";
 import { homedir } from "node:os";
 import { fileURLToPath } from "node:url";
 import { startStandaloneCompanion } from "../knowledge/cli-lifecycle.mjs";

package/payload/shared/scripts/git-runtime-check.mjs

@@ -1,4 +1,4 @@
-import { Ot as GitVersioning } from "./core-entry-Cfvnjfig.mjs";
+import { Ot as GitVersioning } from "./core-entry-CCOXUWXc.mjs";
 import { i as readDesignGit } from "./design-git-read-BUV0s4TT.mjs";
 import { join, resolve } from "node:path";
 import { mkdir, mkdtemp, rm, writeFile } from "node:fs/promises";

package/payload/shared/scripts/knowledge-snapshot-query.mjs

@@ -1,4 +1,4 @@
-import { D as knowledgeReadResult, E as knowledgeQueryFilters, O as queryKnowledgeSnapshot } from "./core-entry-Cfvnjfig.mjs";
+import { D as knowledgeReadResult, E as knowledgeQueryFilters, O as queryKnowledgeSnapshot } from "./core-entry-CCOXUWXc.mjs";
 import { readFile } from "node:fs/promises";
 import { pathToFileURL } from "node:url";
 //#region packages/companion/src/knowledge-snapshot-query.mjs

package/payload/shared/scripts/{load-C46tLUT4.mjs → load-B1XGiCBF.mjs}

@@ -1,4 +1,4 @@
-import { at as loadDesignBundle, ot as loadDesignKit } from "./core-entry-Cfvnjfig.mjs";
+import { at as loadDesignBundle, ot as loadDesignKit } from "./core-entry-CCOXUWXc.mjs";
 import { resolve } from "node:path";
 //#region packages/studio/src/load.ts
 /** Studio-owned loading of a Kit from its directory source. */

package/payload/shared/scripts/{panel-preview-Cb8owVHh.mjs → panel-preview-DvQDHmsY.mjs}

@@ -1,5 +1,5 @@
-import { vt as compileSchema } from "./core-entry-Cfvnjfig.mjs";
-import { t as loadStudioKit } from "./load-C46tLUT4.mjs";
+import { vt as compileSchema } from "./core-entry-CCOXUWXc.mjs";
+import { t as loadStudioKit } from "./load-B1XGiCBF.mjs";
 import { isAbsolute, resolve } from "node:path";
 //#region packages/studio/src/panel-preview.ts
 /** Definition-only Panel preview preparation; never writes or runs a Bundle. */