合并 origin/master：解 4 处冲突并接入 master 的真思考过程

- 合并 origin/master（领先 355+ 提交），冲突 3 文件 4 块，全部按"两侧语义都保住"处理：
- useHomeProjectCreation.ts：保留我们抽出的 uploadLocalFilesAsAttachments（首页与输入盒共用上传链路），同时接入 master 的建项兜底期限（HOME_CREATION_DEADLINE_MS / resolveRecoverableHomeProjectPath）
- ProjectSupervisorView.tsx（属性）：接入 master 新增的 designReasoningEntries 属性
- ProjectSupervisorView.tsx（渲染）：把我们的工具调用折叠块（按回合锚定、实时兜底到列表末尾）与 master 的每消息思考过程（message.reasoningText 折叠块 + designReasoningEntries 兜底）合并，两者同时生效
- view/project-development/index.tsx：接入 master 新增的图标 import（Settings2 / Shapes）
- 类型检查通过（0 错），未跑测试

.env.example

@@ -1,8 +1,8 @@
 # Server-side OpenAI-compatible LLM endpoint base URL.
-LLM_BASE_URL="https://api.vectorengine.cn/v1"
+LLM_BASE_URL="https://api.tiantoken.com/v1"
 
 # Server-side API key used by the local Vite proxy.
-# Recommended: set `LLM_API_KEY` locally, or use `VECTOR_ENGINE_API_KEY`
+# Recommended: set `LLM_API_KEY` locally, or use `TIANTOKEN_API_KEY`
 # through the Rust api-server proxy.
 # Legacy compatibility: `VITE_LLM_API_KEY` is still supported by the proxy,
 # but it should not be relied on by browser code.
@@ -122,7 +122,7 @@ WECHAT_MINIPROGRAM_MESSAGE_ENCODING_AES_KEY=""
 # Model name for chat completions.
 VITE_LLM_MODEL="gpt-5.4-mini"
 GENARRATIVE_LLM_PROVIDER="openai-compatible"
-GENARRATIVE_LLM_BASE_URL="https://api.vectorengine.cn/v1"
+GENARRATIVE_LLM_BASE_URL="https://api.tiantoken.com/v1"
 GENARRATIVE_LLM_API_KEY=""
 GENARRATIVE_LLM_MODEL="gpt-5.4-mini"
 
@@ -130,10 +130,15 @@ GENARRATIVE_LLM_MODEL="gpt-5.4-mini"
 DASHSCOPE_BASE_URL="https://dashscope.aliyuncs.com/api/v1"
 DASHSCOPE_API_KEY="YOUR_DASHSCOPE_API_KEY"
 
-# VectorEngine LLM and GPT-image-2 / Gemini image generation config.
+# Tiantoken LLM and GPT-image-2 / Gemini image generation config.
+TIANTOKEN_BASE_URL="https://api.tiantoken.com"
+TIANTOKEN_API_KEY=""
+TIANTOKEN_IMAGE_REQUEST_TIMEOUT_MS="1000000"
+
+# VectorEngine is retained for Suno audio generation only.
 VECTOR_ENGINE_BASE_URL="https://api.vectorengine.cn"
 VECTOR_ENGINE_API_KEY=""
-VECTOR_ENGINE_IMAGE_REQUEST_TIMEOUT_MS="1000000"
+VECTOR_ENGINE_AUDIO_REQUEST_TIMEOUT_MS="180000"
 
 # ElevenLabs editor sound-effect generation is server-side only.
 ELEVENLABS_BASE_URL="https://api.elevenlabs.io"

.gitea/workflows/project-ci.yml

@@ -27,9 +27,28 @@ env:
   RUSTC_WRAPPER: ''
   CARGO_BUILD_RUSTC_WRAPPER: ''
 
+# job 声明顺序就是 runner 领取顺序，因此把最长尾的客户端 Rust 门禁排在前面，
+# 让它在最少的等待下占用并发槽位；其余 job 按时长递减排列。
+#
+# 客户端（微信壳 / Expo 移动壳 / Tauri 桌面壳 / AI 游戏创作壳）门禁原先全部串在
+# `Native shell tests` 一个 job 里，实测 18 分 37 秒。现在按门禁组拆成一个分组一个 job：
+# `Native shell tests`（契约 + H5 / 微信 / 移动 / 桌面壳门禁 + 发布构建 smoke）、
+# `AI game creator shell web tests`（typecheck + 壳内测试）、`AI game creator shell
+# Rust shard 1/4` 到 `4/4`（AGC 壳 bin 单测按名单分 4 片）、`AI game creator shell
+# Rust smoke`（agent-run smoke）与 `AI game creator shell Rust crates`（AGC 壳依赖的
+# 共享 / 平台 crate 测试）。各自的命令与拆分前逐一对应，本地
+# `npm run check:native-shells` 仍是同一条串行序列。
+#
+# AGC 壳的 bin 单测（2466 条）按名单分 4 片、一片一个 job：片内保持 `--test-threads=1`
+# （当年线程并行会互相干扰的是进程内后台锁与异步终态），片与片之间靠 job 级并发摊开。
+# 不要改回「一个 job 里多进程并行这几片」：同一容器内它们会争抢共享 HOME、target 与固定
+# 临时路径，实测比整套串行还慢。每个分片 job 都会自校验「片并集等于全集且互斥」。
 jobs:
-  repository-checks:
-    name: Repository checks
+  # AGC 壳自身的 Rust bin 单测分片，4 片各自独立 job 并发执行、片内仍保持
+  # `--test-threads=1`。这里不装 npm 依赖：壳 Rust 门禁只用 cargo 与 node 内建模块，
+  # 也只需要 AGC 壳自己那份锁定依赖。
+  ai-game-creator-shell-rust-shard-1:
+    name: AI game creator shell Rust shard 1/4
     runs-on: genarrative-ci
     steps:
       - name: Checkout full history from Gitea
@@ -41,76 +60,228 @@ jobs:
       - name: Validate preinstalled CI job image and sandbox
         run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
 
-      - name: Resolve comparison base
+      - name: Prepare AI game creator shell Rust dependencies
         shell: bash
         run: |
           set -euo pipefail
-          base_ref="$(node -e '
-            const fs = require("node:fs");
-            const event = JSON.parse(fs.readFileSync(process.env.GITHUB_EVENT_PATH, "utf8"));
-            process.stdout.write(event.pull_request?.base?.sha ?? event.before ?? "");
-          ')"
-          if [[ -n "${base_ref}" && ! "${base_ref}" =~ ^0+$ ]]; then
-            git cat-file -e "${base_ref}^{commit}" 2>/dev/null || {
-              echo "comparison base commit is unavailable: ${base_ref}" >&2
+          # AGC 壳有独立 Cargo.lock，其 path 依赖已含 platform-llm、platform-agent、
+          # agent-runtime-core 与 shared-contracts，因此只锁这一份 manifest 就能覆盖壳测试
+          # 与 smoke 的全部第三方依赖；server-rs 那次预热归 crate 级 job，不在这里重复。
+          for attempt in $(seq 1 5); do
+            if cargo fetch --locked \
+              --target x86_64-unknown-linux-gnu \
+              --manifest-path apps/ai-game-creator-shell/src-tauri/Cargo.toml; then
+              break
+            fi
+            if [[ "${attempt}" -eq 5 ]]; then
+              echo 'AI game creator shell Cargo dependency fetch failed after 5 attempts.' >&2
               exit 1
-            }
-          else
-            base_ref="$(git merge-base HEAD origin/master 2>/dev/null || git rev-parse HEAD)"
-          fi
-          resolved_base_ref="$(git rev-parse --verify "${base_ref}^{commit}" 2>/dev/null || true)"
-          head_ref="$(git rev-parse HEAD)"
-          if [[ "${resolved_base_ref}" == "${head_ref}" ]]; then
-            resolved_base_ref="$(git rev-parse --verify HEAD^ 2>/dev/null || true)"
-          fi
-          if [[ -z "${resolved_base_ref}" ]]; then
-            echo 'comparison base must resolve to a commit distinct from HEAD.' >&2
-            exit 1
-          fi
-          base_ref="${resolved_base_ref}"
-          if [[ "${GITHUB_EVENT_NAME:-}" == 'pull_request' ]] \
-            && ! git merge-base --is-ancestor "${base_ref}" HEAD; then
-            echo 'pull request head does not contain the latest base commit; update the branch and rerun CI.' >&2
-            exit 1
-          fi
-          echo "SPACETIME_SCHEMA_BASE_REF=${base_ref}" >> "${GITHUB_ENV}"
+            fi
+            sleep $((attempt * 2))
+          done
 
-      - name: Install npm dependencies
-        run: bash scripts/ci-npm-ci-with-retry.sh
+      - name: Run AI game creator shell Rust shard 1/4
+        run: npm run check:native-shells:agc-rust-shard-1
 
-      - name: Run repository checks
-        run: npm run check:repository-ci
-
-  frontend-tests:
-    name: Frontend tests
+  ai-game-creator-shell-rust-shard-2:
+    name: AI game creator shell Rust shard 2/4
     runs-on: genarrative-ci
     steps:
-      - name: Checkout source from Gitea
+      - name: Checkout full history from Gitea
         env:
-          GENARRATIVE_GITEA_FETCH_DEPTH: '1'
+          GENARRATIVE_GITEA_FETCH_DEPTH: '0'
           GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
         run: genarrative-gitea-checkout
 
       - name: Validate preinstalled CI job image and sandbox
         run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
 
-      - name: Install npm dependencies
-        run: bash scripts/ci-npm-ci-with-retry.sh
+      - name: Prepare AI game creator shell Rust dependencies
+        shell: bash
+        run: |
+          set -euo pipefail
+          for attempt in $(seq 1 5); do
+            if cargo fetch --locked \
+              --target x86_64-unknown-linux-gnu \
+              --manifest-path apps/ai-game-creator-shell/src-tauri/Cargo.toml; then
+              break
+            fi
+            if [[ "${attempt}" -eq 5 ]]; then
+              echo 'AI game creator shell Cargo dependency fetch failed after 5 attempts.' >&2
+              exit 1
+            fi
+            sleep $((attempt * 2))
+          done
 
-      - name: Run frontend and script tests
-        run: npm run test
+      - name: Run AI game creator shell Rust shard 2/4
+        run: npm run check:native-shells:agc-rust-shard-2
 
-      - name: Run BgFilter worker smoke harness tests
-        run: npm run bgfilter-worker:smoke-test
+  ai-game-creator-shell-rust-shard-3:
+    name: AI game creator shell Rust shard 3/4
+    runs-on: genarrative-ci
+    steps:
+      - name: Checkout full history from Gitea
+        env:
+          GENARRATIVE_GITEA_FETCH_DEPTH: '0'
+          GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
+        run: genarrative-gitea-checkout
 
-      - name: Validate production health patrol behavior
-        run: npm run check:production-health-patrol
+      - name: Validate preinstalled CI job image and sandbox
+        run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
 
-      - name: Validate production API release behavior
-        run: npm run check:production-api-release
+      - name: Prepare AI game creator shell Rust dependencies
+        shell: bash
+        run: |
+          set -euo pipefail
+          for attempt in $(seq 1 5); do
+            if cargo fetch --locked \
+              --target x86_64-unknown-linux-gnu \
+              --manifest-path apps/ai-game-creator-shell/src-tauri/Cargo.toml; then
+              break
+            fi
+            if [[ "${attempt}" -eq 5 ]]; then
+              echo 'AI game creator shell Cargo dependency fetch failed after 5 attempts.' >&2
+              exit 1
+            fi
+            sleep $((attempt * 2))
+          done
 
-      - name: Validate production API deploy behavior
-        run: npm run check:production-api-deploy
+      - name: Run AI game creator shell Rust shard 3/4
+        run: npm run check:native-shells:agc-rust-shard-3
+
+  ai-game-creator-shell-rust-shard-4:
+    name: AI game creator shell Rust shard 4/4
+    runs-on: genarrative-ci
+    steps:
+      - name: Checkout full history from Gitea
+        env:
+          GENARRATIVE_GITEA_FETCH_DEPTH: '0'
+          GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
+        run: genarrative-gitea-checkout
+
+      - name: Validate preinstalled CI job image and sandbox
+        run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
+
+      - name: Prepare AI game creator shell Rust dependencies
+        shell: bash
+        run: |
+          set -euo pipefail
+          for attempt in $(seq 1 5); do
+            if cargo fetch --locked \
+              --target x86_64-unknown-linux-gnu \
+              --manifest-path apps/ai-game-creator-shell/src-tauri/Cargo.toml; then
+              break
+            fi
+            if [[ "${attempt}" -eq 5 ]]; then
+              echo 'AI game creator shell Cargo dependency fetch failed after 5 attempts.' >&2
+              exit 1
+            fi
+            sleep $((attempt * 2))
+          done
+
+      - name: Run AI game creator shell Rust shard 4/4
+        run: npm run check:native-shells:agc-rust-shard-4
+
+  # agent-run smoke 会 spawn `cargo run`（走壳自己的 manifest），同样不装 npm 依赖，
+  # 单独一个 job，免得把已经压到 4 分钟级的片 job 拖长。
+  ai-game-creator-shell-rust-smoke:
+    name: AI game creator shell Rust smoke
+    runs-on: genarrative-ci
+    steps:
+      - name: Checkout full history from Gitea
+        env:
+          GENARRATIVE_GITEA_FETCH_DEPTH: '0'
+          GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
+        run: genarrative-gitea-checkout
+
+      - name: Validate preinstalled CI job image and sandbox
+        run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
+
+      - name: Prepare AI game creator shell Rust dependencies
+        shell: bash
+        run: |
+          set -euo pipefail
+          for attempt in $(seq 1 5); do
+            if cargo fetch --locked \
+              --target x86_64-unknown-linux-gnu \
+              --manifest-path apps/ai-game-creator-shell/src-tauri/Cargo.toml; then
+              break
+            fi
+            if [[ "${attempt}" -eq 5 ]]; then
+              echo 'AI game creator shell Cargo dependency fetch failed after 5 attempts.' >&2
+              exit 1
+            fi
+            sleep $((attempt * 2))
+          done
+
+      - name: Run AI game creator shell agent-run smoke
+        run: npm run check:native-shells:agc-rust-smoke
+
+  # AGC 壳依赖的共享 / 平台 crate 测试用的是 server-rs workspace 与两个无锁独立 crate
+  # 的 manifest，属另一套依赖图，因此单独一个 job 预热、单独跑。
+  ai-game-creator-shell-rust-crates:
+    name: AI game creator shell Rust crates
+    runs-on: genarrative-ci
+    steps:
+      - name: Checkout full history from Gitea
+        env:
+          GENARRATIVE_GITEA_FETCH_DEPTH: '0'
+          GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
+        run: genarrative-gitea-checkout
+
+      - name: Validate preinstalled CI job image and sandbox
+        run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
+
+      - name: Prepare server-rs Rust dependencies
+        shell: bash
+        run: |
+          set -euo pipefail
+          for attempt in $(seq 1 5); do
+            if cargo fetch --locked \
+              --target x86_64-unknown-linux-gnu \
+              --manifest-path server-rs/Cargo.toml; then
+              break
+            fi
+            if [[ "${attempt}" -eq 5 ]]; then
+              echo 'server-rs Cargo dependency fetch failed after 5 attempts.' >&2
+              exit 1
+            fi
+            sleep $((attempt * 2))
+          done
+
+      - name: Prepare standalone Rust crate dependencies
+        shell: bash
+        run: |
+          set -euo pipefail
+          # agent-runtime-core / agent-runtime-orchestration 被 server-rs/Cargo.toml 的
+          # exclude 排除，不参与上面的 workspace 锁文件，因此上面那次锁定 fetch 覆盖不到它们；
+          # 而 `npm run ai-game-creator-shell:check:rust:crates` 会用
+          # `cargo test --manifest-path` 单独跑这两个 crate。不在这里预热的话，这两条测试
+          # 会在测试阶段自己 `Updating crates.io index`，crates.io 一抖动整条 job 就红
+          # （见 #327 / PR #316 run 1950）。
+          # 两个 crate 都没有提交 Cargo.lock，所以这里只能做不带锁标志的 fetch：
+          # 加锁标志会因为缺少锁文件直接失败。生成的 Cargo.lock 落在两个 crate 目录内，
+          # 已被各自的 .gitignore 忽略，只留在容器里；随后的测试阶段因此能用锁定版本
+          # 解析，不再触碰 registry index。
+          for manifest_path in \
+            server-rs/crates/agent-runtime-core/Cargo.toml \
+            server-rs/crates/agent-runtime-orchestration/Cargo.toml; do
+            for attempt in $(seq 1 5); do
+              if cargo fetch \
+                --target x86_64-unknown-linux-gnu \
+                --manifest-path "${manifest_path}"; then
+                break
+              fi
+              if [[ "${attempt}" -eq 5 ]]; then
+                echo "standalone crate dependency fetch failed after 5 attempts: ${manifest_path}" >&2
+                exit 1
+              fi
+              sleep $((attempt * 2))
+            done
+          done
+
+      - name: Run AI game creator shell shared crate gates
+        run: npm run check:native-shells:agc-rust-crates
 
   backend-tests:
     name: Backend tests
@@ -194,6 +365,8 @@ jobs:
       - name: Check SpacetimeDB module
         run: cargo check --locked -p spacetime-module --manifest-path server-rs/Cargo.toml
 
+  # 客户端的壳级与契约门禁：静态契约断言、H5 / 微信 / 移动 / 桌面壳运行时门禁，
+  # 以及依赖发布产物的构建 smoke。
   native-shell-tests:
     name: Native shell tests
     runs-on: genarrative-ci
@@ -215,7 +388,6 @@ jobs:
         run: |
           set -euo pipefail
           for manifest_path in \
-            server-rs/Cargo.toml \
             apps/desktop-shell/src-tauri/Cargo.toml \
             apps/ai-game-creator-shell/src-tauri/Cargo.toml; do
             for attempt in $(seq 1 5); do
@@ -232,39 +404,118 @@ jobs:
             done
           done
 
-      - name: Prepare standalone Rust crate dependencies
-        shell: bash
-        run: |
-          set -euo pipefail
-          # agent-runtime-core / agent-runtime-orchestration 被 server-rs/Cargo.toml 的
-          # exclude 排除，不参与上面的 workspace 锁文件，因此上面那次锁定 fetch 覆盖不到它们；
-          # 而 check:native-shells 会经 agent-runtime-*:check 用 `cargo test --manifest-path`
-          # 单独跑这两个 crate。不在这里预热的话，这两条测试会在测试阶段自己
-          # `Updating crates.io index`，crates.io 一抖动整条 native shell 作业就红
-          # （见 #327 / PR #316 run 1950）。
-          # 两个 crate 都没有提交 Cargo.lock，所以这里只能做不带锁标志的 fetch：
-          # 加锁标志会因为缺少锁文件直接失败。生成的 Cargo.lock 落在两个 crate 目录内，
-          # 已被各自的 .gitignore 忽略，只留在容器里；随后的测试阶段因此能用锁定版本
-          # 解析，不再触碰 registry index。
-          for manifest_path in \
-            server-rs/crates/agent-runtime-core/Cargo.toml \
-            server-rs/crates/agent-runtime-orchestration/Cargo.toml; do
-            for attempt in $(seq 1 5); do
-              if cargo fetch \
-                --target x86_64-unknown-linux-gnu \
-                --manifest-path "${manifest_path}"; then
-                break
-              fi
-              if [[ "${attempt}" -eq 5 ]]; then
-                echo "standalone crate dependency fetch failed after 5 attempts: ${manifest_path}" >&2
-                exit 1
-              fi
-              sleep $((attempt * 2))
-            done
-          done
+      - name: Run native shell contract gates
+        run: npm run check:native-shells:contract
 
       - name: Run native shell gates
-        run: npm run check:native-shells
+        run: npm run check:native-shells:shells
+
+      - name: Run native shell release build smoke
+        run: npm run check:native-shells:release
 
       - name: Ensure native lockfiles are unchanged
         run: git diff --exit-code -- apps/desktop-shell/src-tauri/Cargo.lock apps/ai-game-creator-shell/src-tauri/Cargo.lock
+
+  frontend-tests:
+    name: Frontend tests
+    runs-on: genarrative-ci
+    steps:
+      - name: Checkout source from Gitea
+        env:
+          GENARRATIVE_GITEA_FETCH_DEPTH: '1'
+          GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
+        run: genarrative-gitea-checkout
+
+      - name: Validate preinstalled CI job image and sandbox
+        run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
+
+      - name: Install npm dependencies
+        run: bash scripts/ci-npm-ci-with-retry.sh
+
+      - name: Run frontend and script tests
+        run: npm run test
+
+      - name: Run BgFilter worker smoke harness tests
+        run: npm run bgfilter-worker:smoke-test
+
+      - name: Validate production health patrol behavior
+        run: npm run check:production-health-patrol
+
+      - name: Validate production API release behavior
+        run: npm run check:production-api-release
+
+      - name: Validate production API deploy behavior
+        run: npm run check:production-api-deploy
+
+  repository-checks:
+    name: Repository checks
+    runs-on: genarrative-ci
+    steps:
+      - name: Checkout full history from Gitea
+        env:
+          GENARRATIVE_GITEA_FETCH_DEPTH: '0'
+          GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
+        run: genarrative-gitea-checkout
+
+      - name: Validate preinstalled CI job image and sandbox
+        run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
+
+      - name: Resolve comparison base
+        shell: bash
+        run: |
+          set -euo pipefail
+          base_ref="$(node -e '
+            const fs = require("node:fs");
+            const event = JSON.parse(fs.readFileSync(process.env.GITHUB_EVENT_PATH, "utf8"));
+            process.stdout.write(event.pull_request?.base?.sha ?? event.before ?? "");
+          ')"
+          if [[ -n "${base_ref}" && ! "${base_ref}" =~ ^0+$ ]]; then
+            git cat-file -e "${base_ref}^{commit}" 2>/dev/null || {
+              echo "comparison base commit is unavailable: ${base_ref}" >&2
+              exit 1
+            }
+          else
+            base_ref="$(git merge-base HEAD origin/master 2>/dev/null || git rev-parse HEAD)"
+          fi
+          resolved_base_ref="$(git rev-parse --verify "${base_ref}^{commit}" 2>/dev/null || true)"
+          head_ref="$(git rev-parse HEAD)"
+          if [[ "${resolved_base_ref}" == "${head_ref}" ]]; then
+            resolved_base_ref="$(git rev-parse --verify HEAD^ 2>/dev/null || true)"
+          fi
+          if [[ -z "${resolved_base_ref}" ]]; then
+            echo 'comparison base must resolve to a commit distinct from HEAD.' >&2
+            exit 1
+          fi
+          base_ref="${resolved_base_ref}"
+          if [[ "${GITHUB_EVENT_NAME:-}" == 'pull_request' ]] \
+            && ! git merge-base --is-ancestor "${base_ref}" HEAD; then
+            echo 'pull request head does not contain the latest base commit; update the branch and rerun CI.' >&2
+            exit 1
+          fi
+          echo "SPACETIME_SCHEMA_BASE_REF=${base_ref}" >> "${GITHUB_ENV}"
+
+      - name: Install npm dependencies
+        run: bash scripts/ci-npm-ci-with-retry.sh
+
+      - name: Run repository checks
+        run: npm run check:repository-ci
+
+  # 客户端的 AI 游戏创作壳前端门禁：typecheck 与壳内测试，不触碰 Cargo。
+  ai-game-creator-shell-web-tests:
+    name: AI game creator shell web tests
+    runs-on: genarrative-ci
+    steps:
+      - name: Checkout full history from Gitea
+        env:
+          GENARRATIVE_GITEA_FETCH_DEPTH: '0'
+          GENARRATIVE_GITEA_TOKEN: ${{ github.token }}
+        run: genarrative-gitea-checkout
+
+      - name: Validate preinstalled CI job image and sandbox
+        run: GENARRATIVE_GITEA_CI_CHECK_RUNTIME=1 bash scripts/check-gitea-ci-job-image.sh
+
+      - name: Install npm dependencies
+        run: bash scripts/ci-npm-ci-with-retry.sh
+
+      - name: Run AI game creator shell web gates
+        run: npm run check:native-shells:agc-web

apps/ai-game-creator-shell/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@genarrative/ai-game-creator-shell",
   "private": true,
-  "version": "0.1.27",
+  "version": "0.1.29",
   "type": "module",
   "scripts": {
     "dev": "node scripts/start-tauri-dev.mjs",

apps/ai-game-creator-shell/scripts/check-config.mjs

@@ -108,9 +108,16 @@ const rustSharedContractSource = fs.readFileSync(
 );
 const allowedUncalledTauriCommands = [
   'append_direct_project_conversation_message',
+  // TODO: Remove the retired binding command after the legacy runtime path is removed.
+  'bind_components',
   'chat_with_game_creator_agent',
   'check_ui_editor_font_glyph_coverage',
   'create_ui_design_resource',
+  // 图片类生成的同步变体：GUI 已改为 `start_local_project_asset_generation` + 项目内任务账本
+  // （提交即返回、后台生成）。这条命令**没有生产调用方**，只有 Rust 集成测试
+  // （`src/tests/project.rs`）与 `commands.rs` 单测在调；待后续批次删除，或改为转调
+  // `start_local_project_asset_generation`。
+  'generate_local_project_asset',
   'open_game_creator_launcher_window',
   'open_game_creator_workspace_window',
   'read_direct_project_conversation',
@@ -1288,7 +1295,7 @@ for (const requiredSource of [
   }
 }
 
-if (tauriConfig.productName !== 'Genarrative AI Game Creator') {
+if (tauriConfig.productName !== '陶泥儿') {
   throw new Error('AI game creator shell productName drifted');
 }
 
@@ -1300,19 +1307,19 @@ const expectedBundledDesignAgentResources = {
   'design-agent': 'design-agent',
 };
 const expectedBundledWindowsResources = {
-  'resources/codex/win-x64/bin/codex.exe': 'codex/win-x64/bin/codex.exe',
+  'resources/codex/win-x64/bin/codex.exe': 'coding-agent/win-x64/bin/codex.exe',
   'resources/codex/win-x64/bin/codex-code-mode-host.exe':
-    'codex/win-x64/bin/codex-code-mode-host.exe',
+    'coding-agent/win-x64/bin/codex-code-mode-host.exe',
   'resources/codex/win-x64/codex-path/rg.exe':
-    'codex/win-x64/codex-path/rg.exe',
+    'coding-agent/win-x64/codex-path/rg.exe',
   'resources/codex/win-x64/codex-resources/codex-command-runner.exe':
-    'codex/win-x64/codex-resources/codex-command-runner.exe',
+    'coding-agent/win-x64/codex-resources/codex-command-runner.exe',
   'resources/codex/win-x64/codex-resources/codex-windows-sandbox-setup.exe':
-    'codex/win-x64/codex-resources/codex-windows-sandbox-setup.exe',
+    'coding-agent/win-x64/codex-resources/codex-windows-sandbox-setup.exe',
   'resources/codex/win-x64/codex-package.json':
-    'codex/win-x64/codex-package.json',
-  'resources/codex/win-x64/NOTICE.md': 'codex/win-x64/NOTICE.md',
-  'resources/codex/win-x64/manifest.json': 'codex/win-x64/manifest.json',
+    'coding-agent/win-x64/codex-package.json',
+  'resources/codex/win-x64/NOTICE.md': 'coding-agent/win-x64/NOTICE.md',
+  'resources/codex/win-x64/manifest.json': 'coding-agent/win-x64/manifest.json',
   'resources/plugins': 'plugins',
 };
 assert.deepEqual(

apps/ai-game-creator-shell/scripts/dev-port.mjs

@@ -9,6 +9,10 @@ import {
 const agcDevHost = '127.0.0.1';
 const legacyAgcDevPort = 3080;
 const agcVitePortEnvKey = 'GENARRATIVE_AGC_VITE_PORT';
+const agcAdminWebHost = '127.0.0.1';
+const legacyAgcAdminWebPort = 3102;
+// 与 scripts/dev.mjs 的后台 Web 端口配置保持同一环境变量名。
+const agcAdminWebPortEnvKey = 'ADMIN_WEB_PORT';
 
 function readConfiguredAgcDevPort(env = process.env) {
   const rawPort = String(env[agcVitePortEnvKey] ?? '').trim();
@@ -99,13 +103,100 @@ function withAgcDevEndpointEnv(endpoint, env = process.env) {
   };
 }
 
+function readConfiguredAgcAdminWebPort(env = process.env) {
+  const rawPort = String(env[agcAdminWebPortEnvKey] ?? '').trim();
+  if (!rawPort) {
+    return null;
+  }
+
+  const port = normalizePort(rawPort, -1);
+  if (port < 1024) {
+    throw new Error(`${agcAdminWebPortEnvKey} 必须是 1024-65535 的有效端口`);
+  }
+  return port;
+}
+
+function createAgcAdminWebEndpoint(port, portRange = null) {
+  const origin = `http://${agcAdminWebHost}:${port}`;
+  return {
+    host: agcAdminWebHost,
+    port,
+    origin,
+    basePath: '/admin/',
+    url: `${origin}/admin/`,
+    portRange,
+  };
+}
+
+// AGC 开发态的后台 Web 与 `npm run dev` 的后台 Vite 共用同一套优先端口约定：
+// Linux 取当前用户端口段的 `start + 3` 槽位，非 Linux 保留 `3102` 兼容首选并允许统一漂移。
+async function resolveAgcAdminWebEndpoint({
+  env = process.env,
+  platform = process.platform,
+  strictConfigured = false,
+  reservedPorts = [],
+  reservePortRange = reserveLinuxDevPortRange,
+  findPort = findAvailablePort,
+} = {}) {
+  const configuredPort = readConfiguredAgcAdminWebPort(env);
+  let portRange = null;
+  let preferredPort = configuredPort ?? legacyAgcAdminWebPort;
+
+  if (platform === 'linux') {
+    const allocation = await reservePortRange({ env });
+    if (!allocation?.range) {
+      throw new Error('无法取得当前 Linux 用户的 dev 端口段');
+    }
+    portRange = allocation.range;
+    const mappedAdminWebPort = mapDevPortsToPortRange(portRange)?.adminWebPort;
+    if (!Number.isInteger(mappedAdminWebPort)) {
+      throw new Error(
+        `当前 Linux dev 端口段 ${portRange.label} 缺少后台 Web 槽位；请先迁移为至少 6 个端口且不与其它用户重叠的端口段`,
+      );
+    }
+    preferredPort = configuredPort ?? mappedAdminWebPort;
+  }
+
+  const reservedPortSet = new Set(
+    reservedPorts.filter((value) => Number.isInteger(value) && value > 0),
+  );
+  const port = await findPort({
+    host: agcAdminWebHost,
+    preferredPort,
+    portRange,
+    reservedPorts: reservedPortSet,
+    strict: strictConfigured && configuredPort != null,
+  });
+  console.log(
+    formatPortDecision({
+      name: 'ai-game-creator-shell-admin-web',
+      host: agcAdminWebHost,
+      preferredPort,
+      resolvedPort: port,
+    }),
+  );
+  if (portRange) {
+    console.log(
+      `[ai-game-creator-shell] admin-web port-range: ${portRange.label}`,
+    );
+  }
+
+  return createAgcAdminWebEndpoint(port, portRange);
+}
+
 export {
+  agcAdminWebHost,
+  agcAdminWebPortEnvKey,
   agcDevHost,
   agcVitePortEnvKey,
+  createAgcAdminWebEndpoint,
   createAgcDevEndpoint,
+  legacyAgcAdminWebPort,
   legacyAgcDevPort,
   readAgcDevEndpoint,
+  readConfiguredAgcAdminWebPort,
   readConfiguredAgcDevPort,
+  resolveAgcAdminWebEndpoint,
   resolveAgcDevEndpoint,
   withAgcDevEndpointEnv,
 };

apps/ai-game-creator-shell/scripts/run-rust-shell-test-shards.mjs

@@ -0,0 +1,461 @@
+#!/usr/bin/env node
+// AGC 壳 Rust 套件的分片运行器。
+//
+// 背景：AGC 壳的 Rust 单测集中在 src/main.rs 的 bin target（实测 2466 条），仓库口径用
+// `--test-threads=1` 跑，理由是并行调度会让共享 Agent Runtime 后台锁与异步终态的用例在
+// **同一进程内**互相干扰（见 development-workflow 的 Tauri suite 单线程口径）。代价是
+// 整套用例串行跑满 507 秒，占掉 CI 上 `AI game creator shell Rust tests` job 的大头。
+//
+// 这里保留「片内串行」的既有口径，只把用例集合切成 N 片：
+// - CI 用 `--shard-index=<i>` 让**每个 job 只跑一片**，靠多个 job 并发把整套用例摊开；
+// - 本地不传 `--shard-index` 时把 N 片放进 N 个**独立进程**并行（--concurrency 可调），
+//   保留一条命令跑全量的入口。
+// 片并集必须等于全集、且不得重复，数量不符即失败，防止分片规则改动后静默漏跑；该校验
+// 与「只跑一片」无关，因此在每个 job 上都会执行。
+//
+// 注意：同一容器内多进程并行这套用例（共享 HOME、target、固定临时路径）实测会互相拖慢，
+// 比串行还慢，所以 CI 走「一个 job 一片」而不是单 job 内并行。
+//
+// 用法：
+//   node scripts/run-rust-shell-test-shards.mjs --shards=4 --shard-index=2   # CI：只跑第 2 片
+//   node scripts/run-rust-shell-test-shards.mjs [--shards=4] [--concurrency=4]  # 本地：全量
+//   node scripts/run-rust-shell-test-shards.mjs --manifest=<Cargo.toml> --target-kind=lib --no-locked
+//
+// 参数：
+//   --shards=<n>        分片数，默认 4
+//   --shard-index=<i>   只跑第 i 片（1..shards）；不传则跑全部分片
+//   --concurrency=<n>   同时运行的片数，默认等于分片数；--shard-index 时恒为 1
+//   --manifest=<path>   Cargo.toml，默认 ../src-tauri/Cargo.toml（相对本脚本）
+//   --target-kind=<k>   bin | lib，默认 bin（本地自测小 crate 时用 lib）
+//   --bin=<name>        bin target 名，默认 genarrative-ai-game-creator-shell
+//   --package=<name>    target-kind=lib 时要跑的包名（配合 lib 目标使用）
+//   --no-locked         传给 cargo 时不带 --locked（只对没有提交 Cargo.lock 的 crate 需要）
+//   --shard-tmp-root=<path>  片专属 TMPDIR 的父目录，默认 <系统临时目录>/agc-rust-shards
+
+import { spawn } from 'node:child_process';
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const scriptDirectory = path.dirname(fileURLToPath(import.meta.url));
+const shellRoot = path.resolve(scriptDirectory, '..');
+
+function fail(message) {
+  console.error(`[rust-shards] ${message}`);
+  process.exit(1);
+}
+
+function parsePositiveInteger(name, rawValue) {
+  if (!/^[1-9][0-9]*$/.test(rawValue)) {
+    fail(`${name} must be a positive integer, received: ${rawValue}`);
+  }
+
+  return Number(rawValue);
+}
+
+const options = {
+  shards: 4,
+  shardIndex: undefined,
+  concurrency: undefined,
+  manifestPath: path.join(shellRoot, 'src-tauri', 'Cargo.toml'),
+  targetKind: 'bin',
+  binName: 'genarrative-ai-game-creator-shell',
+  packageName: undefined,
+  locked: true,
+  shardTmpRoot: path.join(os.tmpdir(), 'agc-rust-shards'),
+};
+
+for (const rawArgument of process.argv.slice(2)) {
+  if (rawArgument === '--no-locked') {
+    options.locked = false;
+    continue;
+  }
+
+  const separatorIndex = rawArgument.indexOf('=');
+  if (!rawArgument.startsWith('--') || separatorIndex === -1) {
+    fail(`unexpected argument: ${rawArgument}`);
+  }
+
+  const name = rawArgument.slice(2, separatorIndex);
+  const value = rawArgument.slice(separatorIndex + 1);
+  switch (name) {
+    case 'shards':
+      options.shards = parsePositiveInteger('--shards', value);
+      break;
+    case 'shard-index':
+      options.shardIndex = parsePositiveInteger('--shard-index', value);
+      break;
+    case 'concurrency':
+      options.concurrency = parsePositiveInteger('--concurrency', value);
+      break;
+    case 'manifest':
+      options.manifestPath = path.resolve(process.cwd(), value);
+      break;
+    case 'target-kind':
+      if (value !== 'bin' && value !== 'lib') {
+        fail(`--target-kind must be bin or lib, received: ${value}`);
+      }
+      options.targetKind = value;
+      break;
+    case 'bin':
+      options.binName = value;
+      break;
+    case 'package':
+      options.packageName = value;
+      break;
+    case 'shard-tmp-root':
+      options.shardTmpRoot = path.resolve(process.cwd(), value);
+      break;
+    default:
+      fail(`unexpected argument: ${rawArgument}`);
+  }
+}
+
+if (!fs.existsSync(options.manifestPath)) {
+  fail(`manifest does not exist: ${options.manifestPath}`);
+}
+
+if (options.shardIndex !== undefined && options.shardIndex > options.shards) {
+  fail(
+    `--shard-index (${options.shardIndex}) must be within --shards (${options.shards})`,
+  );
+}
+
+const concurrency = options.concurrency ?? options.shards;
+const crateRoot = path.dirname(options.manifestPath);
+
+function buildCargoArguments(target) {
+  const cargoArguments = ['test'];
+  if (options.locked) {
+    cargoArguments.push('--locked');
+  }
+  cargoArguments.push('--manifest-path', options.manifestPath);
+  if (target.kind === 'lib') {
+    if (options.packageName !== undefined) {
+      cargoArguments.push('-p', options.packageName);
+    }
+    cargoArguments.push('--lib');
+    return cargoArguments;
+  }
+  cargoArguments.push('--bin', target.name);
+  return cargoArguments;
+}
+
+function formatDuration(milliseconds) {
+  return `${(milliseconds / 1000).toFixed(1)}s`;
+}
+
+// 编译一次，直接拿到测试可执行文件：后续每片都运行同一个二进制，不再各自调用 cargo，
+// 免得 N 个 cargo 去争 package cache 与 target 目录锁。
+function resolveTestExecutable() {
+  return new Promise((resolve, reject) => {
+    const cargoArguments = buildCargoArguments({
+      kind: options.targetKind,
+      name: options.binName,
+    });
+    cargoArguments.push('--no-run', '--message-format=json');
+    console.log(`[rust-shards] cargo ${cargoArguments.join(' ')}`);
+
+    const child = spawn('cargo', cargoArguments, {
+      cwd: crateRoot,
+      env: process.env,
+      stdio: ['ignore', 'pipe', 'inherit'],
+    });
+
+    let buffered = '';
+    const executables = [];
+    child.stdout.setEncoding('utf8');
+    child.stdout.on('data', (chunk) => {
+      buffered += chunk;
+      const lines = buffered.split('\n');
+      buffered = lines.pop() ?? '';
+      for (const line of lines) {
+        if (!line.startsWith('{')) {
+          continue;
+        }
+        let message;
+        try {
+          message = JSON.parse(line);
+        } catch {
+          continue;
+        }
+        if (
+          message.reason === 'compiler-artifact' &&
+          typeof message.executable === 'string'
+        ) {
+          executables.push(message.executable);
+        }
+      }
+    });
+
+    child.on('error', (error) => {
+      reject(new Error(`unable to start cargo: ${error.message}`));
+    });
+    child.on('close', (code) => {
+      if (code !== 0) {
+        reject(
+          new Error(
+            `cargo ${cargoArguments.join(' ')} failed with exit code ${code}`,
+          ),
+        );
+        return;
+      }
+      const uniqueExecutables = [...new Set(executables)];
+      if (uniqueExecutables.length !== 1) {
+        reject(
+          new Error(
+            `expected exactly one test executable for the ${options.targetKind} target, found ${uniqueExecutables.length}: ${uniqueExecutables.join(', ')}`,
+          ),
+        );
+        return;
+      }
+      resolve(uniqueExecutables[0]);
+    });
+  });
+}
+
+function listTestNames(executable) {
+  return new Promise((resolve, reject) => {
+    const child = spawn(executable, ['--list'], {
+      cwd: crateRoot,
+      env: process.env,
+      stdio: ['ignore', 'pipe', 'pipe'],
+    });
+
+    let stdout = '';
+    let stderr = '';
+    child.stdout.setEncoding('utf8');
+    child.stderr.setEncoding('utf8');
+    child.stdout.on('data', (chunk) => {
+      stdout += chunk;
+    });
+    child.stderr.on('data', (chunk) => {
+      stderr += chunk;
+    });
+    child.on('error', (error) => {
+      reject(new Error(`unable to list tests: ${error.message}`));
+    });
+    child.on('close', (code) => {
+      if (code !== 0) {
+        reject(
+          new Error(
+            `listing tests failed with exit code ${code}: ${stderr.trim()}`,
+          ),
+        );
+        return;
+      }
+      const names = [];
+      for (const line of stdout.split('\n')) {
+        const match = /^(.*): test$/.exec(line.trim());
+        if (match !== null && match[1].length > 0) {
+          names.push(match[1]);
+        }
+      }
+      resolve(names);
+    });
+  });
+}
+
+function splitTestNames(testNames, shardCount) {
+  const sortedNames = [...testNames].sort();
+  const shards = Array.from({ length: shardCount }, () => []);
+  sortedNames.forEach((testName, index) => {
+    shards[index % shardCount].push(testName);
+  });
+  return shards;
+}
+
+function assertShardsCoverEveryTest(testNames, shards) {
+  const flattened = shards.flat();
+  if (flattened.length !== testNames.length) {
+    fail(
+      `shard split covered ${flattened.length} of ${testNames.length} tests; the split rule must be exhaustive`,
+    );
+  }
+  if (new Set(flattened).size !== flattened.length) {
+    fail(
+      'shard split selected the same test more than once; the split rule must be disjoint',
+    );
+  }
+  const missing = testNames.filter((testName) => !flattened.includes(testName));
+  if (missing.length > 0) {
+    fail(
+      `shard split is missing tests, for example: ${missing.slice(0, 5).join(', ')}`,
+    );
+  }
+}
+
+function runShard(executable, shardIndex, shardCount, shardTestNames) {
+  const label = `shard ${shardIndex + 1}/${shardCount}`;
+  const shardTmpDirectory = path.join(
+    options.shardTmpRoot,
+    `shard-${shardIndex + 1}`,
+  );
+  fs.rmSync(shardTmpDirectory, { recursive: true, force: true });
+  fs.mkdirSync(shardTmpDirectory, { recursive: true });
+  const startedAt = Date.now();
+
+  return new Promise((resolve) => {
+    const child = spawn(
+      executable,
+      ['--exact', ...shardTestNames, '--test-threads=1'],
+      {
+        cwd: crateRoot,
+        env: {
+          ...process.env,
+          TMPDIR: shardTmpDirectory,
+          TMP: shardTmpDirectory,
+          TEMP: shardTmpDirectory,
+        },
+        stdio: ['ignore', 'pipe', 'pipe'],
+      },
+    );
+
+    const failureLines = [];
+    let inFailureList = false;
+    let stderr = '';
+    const consumeLine = (rawLine) => {
+      const line = rawLine.replace(/\r$/, '');
+      if (line.includes('failures:')) {
+        inFailureList = true;
+        return;
+      }
+      if (inFailureList) {
+        if (line.trim().length === 0) {
+          inFailureList = false;
+          return;
+        }
+        failureLines.push(line.trim());
+      }
+    };
+
+    child.stdout.setEncoding('utf8');
+    child.stderr.setEncoding('utf8');
+    let stdoutBuffer = '';
+    child.stdout.on('data', (chunk) => {
+      stdoutBuffer += chunk;
+      const lines = stdoutBuffer.split('\n');
+      stdoutBuffer = lines.pop() ?? '';
+      for (const line of lines) {
+        consumeLine(line);
+      }
+    });
+    child.stderr.on('data', (chunk) => {
+      stderr += chunk;
+    });
+
+    child.on('error', (error) => {
+      resolve({
+        label,
+        ok: false,
+        durationMs: Date.now() - startedAt,
+        testCount: shardTestNames.length,
+        failures: [`unable to start test binary: ${error.message}`],
+        stderr,
+      });
+    });
+    child.on('close', (code) => {
+      resolve({
+        label,
+        ok: code === 0,
+        durationMs: Date.now() - startedAt,
+        testCount: shardTestNames.length,
+        failures: failureLines,
+        stderr,
+      });
+    });
+  });
+}
+
+async function runWithConcurrency(shards, runner) {
+  const results = new Array(shards.length);
+  let nextIndex = 0;
+  const workers = Array.from(
+    { length: Math.min(concurrency, shards.length) },
+    async () => {
+      while (nextIndex < shards.length) {
+        const index = nextIndex;
+        nextIndex += 1;
+        results[index] = await runner(shards[index], index);
+      }
+    },
+  );
+  await Promise.all(workers);
+  return results;
+}
+
+async function main() {
+  const executable = await resolveTestExecutable();
+  const testNames = await listTestNames(executable);
+  if (testNames.length === 0) {
+    fail(
+      `no tests discovered in ${options.manifestPath} (${options.targetKind})`,
+    );
+  }
+
+  const shards = splitTestNames(testNames, options.shards);
+  assertShardsCoverEveryTest(testNames, shards);
+
+  const selectedShards =
+    options.shardIndex === undefined
+      ? shards.map((shardTestNames, index) => ({ index, shardTestNames }))
+      : [
+          {
+            index: options.shardIndex - 1,
+            shardTestNames: shards[options.shardIndex - 1],
+          },
+        ];
+
+  if (options.shardIndex === undefined) {
+    console.log(
+      `[rust-shards] ${testNames.length} tests, ${shards.length} shard(s), concurrency ${Math.min(concurrency, shards.length)}`,
+    );
+    for (const { index, shardTestNames } of selectedShards) {
+      console.log(
+        `[rust-shards] shard ${index + 1}/${shards.length}: ${shardTestNames.length} test(s)`,
+      );
+    }
+  } else {
+    console.log(
+      `[rust-shards] ${testNames.length} tests, ${shards.length} shard(s), running shard ${options.shardIndex}/${shards.length} (${selectedShards[0].shardTestNames.length} test(s))`,
+    );
+  }
+
+  const results = await runWithConcurrency(
+    selectedShards,
+    ({ index, shardTestNames }) =>
+      runShard(executable, index, shards.length, shardTestNames),
+  );
+
+  let failed = false;
+  for (const result of results) {
+    if (result.ok) {
+      console.log(
+        `[rust-shards] ${result.label} ok: ${result.testCount} test(s) in ${formatDuration(result.durationMs)}`,
+      );
+      continue;
+    }
+    failed = true;
+    console.error(
+      `[rust-shards] ${result.label} FAILED: ${result.testCount} test(s) in ${formatDuration(result.durationMs)}`,
+    );
+    for (const failure of result.failures) {
+      console.error(`[rust-shards]   ${failure}`);
+    }
+    if (result.stderr.trim().length > 0) {
+      console.error(
+        `[rust-shards]   stderr: ${result.stderr.trim().split('\n').slice(-20).join('\n[rust-shards]   ')}`,
+      );
+    }
+  }
+
+  if (failed) {
+    process.exit(1);
+  }
+  console.log('[rust-shards] OK');
+}
+
+main().catch((error) => {
+  fail(error instanceof Error ? error.message : String(error));
+});

apps/ai-game-creator-shell/scripts/skill-pack-manifest.mjs

@@ -8,6 +8,7 @@ export const SKILL_PACK_SCHEMA_VERSION = 'agc-skill-pack.v1';
 export const EXPECTED_SKILL_NAMES = Object.freeze([
   'agc-browser-playtest',
   'agc-client-projection',
+  'agc-game-production-workflow',
   'agc-project-structure',
   'agc-web-game-development',
   'taonier-art-assets',

apps/ai-game-creator-shell/scripts/start-dev-stack.mjs

@@ -14,12 +14,14 @@ import {
 import {
   agcVitePortEnvKey,
   readAgcDevEndpoint,
+  resolveAgcAdminWebEndpoint,
   resolveAgcDevEndpoint,
   withAgcDevEndpointEnv,
 } from './dev-port.mjs';
 
 const appRoot = fileURLToPath(new URL('..', import.meta.url));
 const repoRoot = resolve(appRoot, '../..');
+const adminWebDir = resolve(repoRoot, 'apps/admin-web');
 const devStackStatePath = resolve(repoRoot, '.app/dev-stack.json');
 const apiServerExePath = resolve(
   repoRoot,
@@ -32,6 +34,8 @@ const backendSpacetimeDataDir = resolve(
   repoRoot,
   'server-rs/.spacetimedb/ai-game-creator/data',
 );
+// 后台 Web 默认跟随 AGC 一起起来，便于联调后台页面；`AGC_DEV_ADMIN_WEB=0` 可关闭。
+const agcDevAdminWebEnvKey = 'AGC_DEV_ADMIN_WEB';
 const npm = process.platform === 'win32' ? 'npm.cmd' : 'npm';
 const childLifecycles = new WeakMap();
 
@@ -84,6 +88,7 @@ function resolveBackendTargetsFromState(
     requireAgcBackend = false,
     expectedDatabase = backendDatabase,
     expectedSpacetimeDataDir = backendSpacetimeDataDir,
+    expectedRepoRoot = repoRoot,
     fallbackApiTarget = defaultApiTarget,
   } = {},
 ) {
@@ -101,7 +106,25 @@ function resolveBackendTargetsFromState(
   const hasMatchingDataDir =
     Boolean(spacetimeDataDir) &&
     spacetimeDataDir === resolve(expectedSpacetimeDataDir);
-  const hasMatchingBackend = hasMatchingDatabase && hasMatchingDataDir;
+  const instanceId =
+    typeof state?.instanceId === 'string' ? state.instanceId.trim() : '';
+  const hasMatchingRepoRoot =
+    typeof state?.repoRoot === 'string' &&
+    resolve(state.repoRoot) === resolve(expectedRepoRoot);
+  const hasMatchingInstance =
+    Boolean(instanceId) &&
+    [apiServer, spacetime, bgfilterWorker]
+      .filter(Boolean)
+      .every(
+        (service) =>
+          service.repoRoot &&
+          resolve(service.repoRoot) === resolve(expectedRepoRoot) &&
+          service.instanceId === instanceId,
+      );
+  const hasMatchingBackend =
+    hasMatchingDatabase &&
+    hasMatchingDataDir &&
+    (!requireAgcBackend || (hasMatchingRepoRoot && hasMatchingInstance));
   const canReuseState = !requireAgcBackend || hasMatchingBackend;
   const apiUrl =
     canReuseState && isActive(apiServer) && apiServer.url
@@ -127,6 +150,8 @@ function resolveBackendTargetsFromState(
     spacetimeDataDir,
     hasMatchingDatabase,
     hasMatchingDataDir,
+    hasMatchingRepoRoot,
+    hasMatchingInstance,
     hasMatchingBackend,
   };
 }
@@ -179,36 +204,122 @@ function urlPort(url) {
   }
 }
 
-// 读取端口当前真正的监听进程身份。返回 null 表示探测本身不可用（例如缺少
-// Get-NetTCPConnection），此时调用方必须退化为旧行为，不能让本地启动直接失败。
+// 端口归属探测脚本。历史实现用 `Get-NetTCPConnection` 取监听进程，而它底层走
+// WMI：实测单端口单次 11.2 秒、再叠加每个 PID 的 `Get-CimInstance` 3.3 秒，
+// 一轮探测约 43 秒，直接把"配套后端就绪"等待拖到分钟级。改用原生
+// `netstat -ano`（约 30 毫秒）取端口 -> PID，再用 .NET `Process` 读进程名和
+// 可执行文件路径（毫秒级）；只有核对 SpacetimeDB `--data-dir` 归属时才按 PID
+// 取命令行，并允许调用方把已知命令行传进来复用。
+const windowsPortOwnerProbeCommand = [
+  '$ErrorActionPreference = "SilentlyContinue"',
+  '$queriedPorts = @()',
+  'foreach ($raw in ($env:GENARRATIVE_QUERY_PORTS -split ",")) {',
+  '  if ($raw -match "^\\d+$") { $queriedPorts += [int]$raw }',
+  '}',
+  '$knownCommandLines = @{}',
+  'if ($env:GENARRATIVE_KNOWN_COMMAND_LINES) {',
+  '  try {',
+  '    foreach ($property in (ConvertFrom-Json $env:GENARRATIVE_KNOWN_COMMAND_LINES).PSObject.Properties) {',
+  '      $knownCommandLines[[int]$property.Name] = [string]$property.Value',
+  '    }',
+  '  } catch { }',
+  '}',
+  '$listenerPidByPort = @{}',
+  'foreach ($line in (netstat -ano -p tcp)) {',
+  '  $fields = @($line -split "\\s+" | Where-Object { $_ })',
+  '  if ($fields.Count -lt 4) { continue }',
+  '  if ($fields[0] -ne "TCP") { continue }',
+  '  # A listening socket always has foreign address 0.0.0.0:0 / [::]:0, which',
+  '  # is locale-independent unlike the localized netstat State column.',
+  '  if ($fields[2] -notmatch ":0$") { continue }',
+  '  $localPort = [int]($fields[1].Split(":")[-1])',
+  '  if ($queriedPorts -notcontains $localPort) { continue }',
+  '  # The PID is the last column; do not hardcode its index.',
+  '  if ($fields[-1] -notmatch "^\\d+$") { continue }',
+  '  $listenerPidByPort[$localPort] = [int]$fields[-1]',
+  '}',
+  '$result = @()',
+  'foreach ($port in ($listenerPidByPort.Keys | Sort-Object)) {',
+  '  $processId = $listenerPidByPort[$port]',
+  '  $name = $null',
+  '  $executablePath = $null',
+  '  $commandLine = $null',
+  '  try {',
+  '    $process = [System.Diagnostics.Process]::GetProcessById($processId)',
+  '    $name = $process.ProcessName + ".exe"',
+  '    try { $executablePath = $process.MainModule.FileName } catch { }',
+  '  } catch { }',
+  '  if ($knownCommandLines.ContainsKey($processId)) {',
+  '    $commandLine = $knownCommandLines[$processId]',
+  '  } elseif (($name -like "spacetime*") -or (-not $executablePath)) {',
+  '    try { $commandLine = (Get-CimInstance Win32_Process -Filter ("ProcessId=" + $processId)).CommandLine } catch { }',
+  '  }',
+  '  $result += [pscustomobject]@{ port = [int]$port; processId = $processId; name = $name; executablePath = $executablePath; commandLine = $commandLine }',
+  '}',
+  'ConvertTo-Json -InputObject @($result) -Compress',
+].join('\n');
+
+// 进程命令行在进程生命周期内不变，但 PID 会被系统复用；按 PID 记 TTL 缓存，
+// 让"等配套后端就绪"的轮询只在首个周期付出 WMI 成本。TTL 取 5 分钟：本轮实测
+// 这台机器上首次 WMI 调用约 18 秒（热调用 3.3 秒），而 PID 在 5 分钟内被复用
+// 成另一个运行本工作树 data dir 的 SpacetimeDB 才能造成误判，概率可忽略。
+// 默认实现才缓存，注入实现（测试）与显式 env 始终重新读取。
+const WINDOWS_COMMAND_LINE_CACHE_TTL_MS = 300_000;
+const windowsPortOwnerCommandLineCache = new Map();
+
+function resolveCommandLineCache({ spawnImpl, env }) {
+  return spawnImpl === spawnSync && env === process.env
+    ? windowsPortOwnerCommandLineCache
+    : new Map();
+}
+
+// 读取端口当前真正的监听进程身份。返回 null 表示探测本身不可用（例如系统缺少
+// netstat），此时调用方必须退化为旧行为，不能让本地启动直接失败。
 function readWindowsPortOwnerIdentities(
   ports,
-  { spawnImpl = spawnSync, env = process.env } = {},
+  {
+    spawnImpl = spawnSync,
+    env = process.env,
+    now = Date.now,
+    commandLineTtlMs = WINDOWS_COMMAND_LINE_CACHE_TTL_MS,
+    commandLineCache = resolveCommandLineCache({ spawnImpl, env }),
+  } = {},
 ) {
   const uniquePorts = [...new Set(ports.filter((port) => port > 0))];
   if (uniquePorts.length === 0) {
     return null;
   }
 
-  const command = [
-    '$ErrorActionPreference = "SilentlyContinue"',
-    '$ports = ($env:GENARRATIVE_QUERY_PORTS -split ",") | Where-Object { $_ }',
-    '$result = @()',
-    'foreach ($port in $ports) {',
-    '  $connection = Get-NetTCPConnection -State Listen -LocalPort ([int]$port) -ErrorAction SilentlyContinue | Select-Object -First 1',
-    '  if (-not $connection) { continue }',
-    '  $owner = Get-CimInstance Win32_Process -Filter ("ProcessId=" + $connection.OwningProcess) -ErrorAction SilentlyContinue',
-    '  $result += [pscustomobject]@{ port = [int]$port; processId = [int]$connection.OwningProcess; name = $owner.Name; executablePath = $owner.ExecutablePath; commandLine = $owner.CommandLine }',
-    '}',
-    'ConvertTo-Json -InputObject @($result) -Compress',
-  ].join('\n');
+  const knownCommandLines = {};
+  for (const [processId, record] of [...commandLineCache]) {
+    if (record && now() - record.at < commandLineTtlMs) {
+      knownCommandLines[processId] = record.commandLine;
+    } else {
+      commandLineCache.delete(processId);
+    }
+  }
+
+  const childEnv = {
+    ...env,
+    GENARRATIVE_QUERY_PORTS: uniquePorts.join(','),
+  };
+  if (Object.keys(knownCommandLines).length > 0) {
+    childEnv.GENARRATIVE_KNOWN_COMMAND_LINES =
+      JSON.stringify(knownCommandLines);
+  }
 
   const result = spawnImpl(
     'powershell.exe',
-    ['-NoProfile', '-ExecutionPolicy', 'Bypass', '-Command', command],
+    [
+      '-NoProfile',
+      '-ExecutionPolicy',
+      'Bypass',
+      '-Command',
+      windowsPortOwnerProbeCommand,
+    ],
     {
       encoding: 'utf8',
-      env: { ...env, GENARRATIVE_QUERY_PORTS: uniquePorts.join(',') },
+      env: childEnv,
       maxBuffer: 8 * 1024 * 1024,
     },
   );
@@ -219,9 +330,22 @@ function readWindowsPortOwnerIdentities(
   const owners = new Map();
   for (const entry of parseWindowsProcessSnapshot(result.stdout)) {
     const port = Number(entry?.port);
-    if (Number.isInteger(port) && port > 0) {
-      owners.set(port, entry);
+    if (!Number.isInteger(port) || port <= 0) {
+      continue;
     }
+    const processId = Number(entry?.processId);
+    if (
+      Number.isInteger(processId) &&
+      processId > 0 &&
+      typeof entry?.commandLine === 'string' &&
+      entry.commandLine
+    ) {
+      commandLineCache.set(processId, {
+        commandLine: entry.commandLine,
+        at: now(),
+      });
+    }
+    owners.set(port, entry);
   }
   return owners;
 }
@@ -858,10 +982,102 @@ async function startVite(apiTarget, endpoint = readAgcDevEndpoint()) {
   );
 }
 
+function readAdminWebEnabled(env = process.env) {
+  return String(env[agcDevAdminWebEnvKey] ?? '').trim() !== '0';
+}
+
+// 后台 Web 与 AGC Vite 一样直接由本启动器持有，不经过 `dev.mjs admin-web`：
+// 后者会整体重写 `.app/dev-stack.json`，把本次配套后端的状态覆盖掉。
+function startAdminWeb(
+  apiUrl,
+  endpoint,
+  { env = process.env, spawnImpl = spawnChild } = {},
+) {
+  return spawnImpl(
+    npm,
+    [
+      '--prefix',
+      '../..',
+      'exec',
+      'vite',
+      '--',
+      '--host',
+      endpoint.host,
+      '--port',
+      String(endpoint.port),
+      '--strictPort',
+    ],
+    {
+      cwd: adminWebDir,
+      env: {
+        ...env,
+        ADMIN_API_TARGET: apiUrl,
+        GENARRATIVE_API_TARGET: apiUrl,
+        GENARRATIVE_API_PORT: String(urlPort(apiUrl) || 8082),
+        ADMIN_WEB_BASE: endpoint.basePath,
+      },
+    },
+  );
+}
+
+function formatStartupSummary({
+  frontendUrl = '',
+  apiUrl = '',
+  adminWebUrl = '',
+  spacetimeUrl = '',
+  bgfilterWorkerUrl = '',
+} = {}) {
+  const segments = [
+    ['前端', frontendUrl],
+    ['后端', apiUrl],
+    ['后台', adminWebUrl],
+    ['数据库', spacetimeUrl],
+    ['bgfilter-worker', bgfilterWorkerUrl],
+  ]
+    .filter(([, value]) => Boolean(value))
+    .map(([label, value]) => `${label} ${value}`);
+  return `[ai-game-creator-shell] 启动汇总: ${segments.join(' | ')}`;
+}
+
+// 后台 Web 是可选联调服务：端口解析或启动失败只告警，不能阻断 AGC 客户端与配套后端。
+async function ensureAdminWeb({
+  apiUrl,
+  reservedPorts = [],
+  env = process.env,
+  enabled = readAdminWebEnabled(env),
+  resolveEndpoint = resolveAgcAdminWebEndpoint,
+  spawnAdminWeb = startAdminWeb,
+  waitForExit = waitForChildTermination,
+  warn = (message) => console.warn(message),
+} = {}) {
+  if (!enabled) {
+    return { endpoint: null, child: null };
+  }
+
+  try {
+    const endpoint = await resolveEndpoint({ env, reservedPorts });
+    const child = spawnAdminWeb(apiUrl, endpoint, { env });
+    waitForExit(child).then((failure) => {
+      warn(
+        `[ai-game-creator-shell] 后台 Web 已退出（${formatChildFailure(failure)}），AGC 继续运行。`,
+      );
+    });
+    return { endpoint, child };
+  } catch (error) {
+    warn(
+      `[ai-game-creator-shell] 后台 Web 未能启动（${
+        error instanceof Error ? error.message : String(error)
+      }），AGC 继续运行。`,
+    );
+    return { endpoint: null, child: null };
+  }
+}
+
 async function main() {
   let backendChild = null;
   let startedBackend = false;
   let viteChild = null;
+  let adminWebChild = null;
   let shutdownSignal = '';
   const signalHandlers = new Map();
 
@@ -884,6 +1100,7 @@ async function main() {
     const handler = () => {
       shutdownSignal = signal;
       stopChild(viteChild, signal);
+      stopChild(adminWebChild, signal);
       stopChild(backendChild, signal);
       // 立刻清扫，避免外层 taskkill /F 抢在 finally 之前把本进程杀掉。
       sweepStartedBackend();
@@ -916,6 +1133,25 @@ async function main() {
       throw new Error(`启动期收到 ${shutdownSignal}，已停止前端服务`);
     }
 
+    const adminWeb = await ensureAdminWeb({
+      apiUrl: backend.targets.apiUrl,
+      // AGC Vite 端口尚未监听，必须显式保留，避免被后台 Web 抢先占用。
+      reservedPorts: [endpoint.port],
+    });
+    adminWebChild = adminWeb.child;
+    if (shutdownSignal) {
+      throw new Error(`启动期收到 ${shutdownSignal}，已停止后台 Web`);
+    }
+    console.log(
+      formatStartupSummary({
+        frontendUrl: endpoint.url,
+        apiUrl: backend.targets.apiUrl,
+        adminWebUrl: adminWeb.endpoint?.url ?? '',
+        spacetimeUrl: backend.targets.spacetimeUrl,
+        bgfilterWorkerUrl: backend.targets.bgfilterWorkerUrl,
+      }),
+    );
+
     const children = [backendChild, viteChild].filter(Boolean);
     if (children.length === 0) {
       return 0;
@@ -925,10 +1161,12 @@ async function main() {
       children.map((child) => waitForChildTermination(child)),
     );
     stopChild(viteChild);
+    stopChild(adminWebChild);
     stopChild(backendChild);
     return failure.type === 'error' || failure.signal ? 1 : (failure.code ?? 0);
   } catch (error) {
     stopChild(viteChild);
+    stopChild(adminWebChild);
     stopChild(backendChild);
     console.error(
       `[ai-game-creator-shell] ${error instanceof Error ? error.message : String(error)}`,
@@ -937,6 +1175,7 @@ async function main() {
   } finally {
     await Promise.all([
       terminateChildTree(viteChild),
+      terminateChildTree(adminWebChild),
       terminateChildTree(backendChild),
     ]);
     sweepStartedBackend();
@@ -954,9 +1193,12 @@ function isDirectModuleExecution() {
 }
 
 export {
+  agcDevAdminWebEnvKey,
+  ensureAdminWeb,
   ensureBackend,
   formatChildFailure,
   formatOwnerLabel,
+  formatStartupSummary,
   isAiGameCreatorServer,
   isBackendReady,
   isDirectModuleExecution,
@@ -964,6 +1206,7 @@ export {
   isWorktreeApiServerOwner,
   isWorktreeSpacetimeOwner,
   preflightExistingVite,
+  readAdminWebEnabled,
   readBackendServiceFailure,
   readChildFailure,
   readExistingViteServer,
@@ -972,6 +1215,7 @@ export {
   resolveBackendTargetsFromState,
   runWindowsTaskkill,
   spawnChild,
+  startAdminWeb,
   stopChild,
   terminateChildTree,
   verifyAgcBackendOwnership,

apps/ai-game-creator-shell/scripts/start-tauri-dev.mjs

@@ -22,7 +22,6 @@ const appRoot = fileURLToPath(new URL('..', import.meta.url));
 const repoRoot = resolve(appRoot, '../..');
 const tauriCliPath = resolve(repoRoot, 'node_modules/@tauri-apps/cli/tauri.js');
 const AGC_DESIGN_DEBUG_ENV = 'GENARRATIVE_AGC_DESIGN_DEBUG';
-const AGC_DESIGN_DEBUG_VITE_ENV = 'VITE_GENARRATIVE_AGC_DESIGN_DEBUG';
 const designDebugEnabled =
   process.env[AGC_DESIGN_DEBUG_ENV]?.trim() === '0' ? '0' : '1';
 
@@ -137,7 +136,6 @@ async function runTauriDev(
       env: {
         ...withAgcDevEndpointEnv(endpoint),
         [AGC_DESIGN_DEBUG_ENV]: designDebugEnabled,
-        [AGC_DESIGN_DEBUG_VITE_ENV]: designDebugEnabled,
       },
     });
     const childResult = waitForCli(child);
@@ -191,7 +189,6 @@ async function prepareFrontendDev(endpoint, { onChild, signal }) {
       cwd: repoRoot,
       env: {
         ...withAgcDevEndpointEnv(endpoint),
-        [AGC_DESIGN_DEBUG_VITE_ENV]: designDebugEnabled,
       },
     },
   );

apps/ai-game-creator-shell/src-tauri/Cargo.lock

@@ -1725,7 +1725,7 @@ dependencies = [
 
 [[package]]
 name = "genarrative-ai-game-creator-shell"
-version = "0.1.27"
+version = "0.1.29"
 dependencies = [
  "agent-runtime-core",
  "axum",

apps/ai-game-creator-shell/src-tauri/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "genarrative-ai-game-creator-shell"
-version = "0.1.27"
+version = "0.1.29"
 edition = "2021"
 publish = false
 

apps/ai-game-creator-shell/src-tauri/design-agent/resources/skills/architecture.md

@@ -74,7 +74,7 @@ description: 写游戏策划案（GDD）系统架构时使用。在顶层设计
 三个接口：**对上**承顶层系统范围表并跑循环覆盖检查；**对内**地图↔职责↔依赖
 三方一致、主数据归属唯一；**对下**目录映射 + MVP 闭环喂系统文档站。
 
-## 四、怎么写（模板即流程，十二节按序）
+## 四、怎么写（模板参考结构，建议按此组织）
 （本节是带写法要领的教学版；实际填写的纯净模板在 templates/architecture.md）
 
 ### 1. 架构定位与目标
@@ -85,7 +85,7 @@ description: 写游戏策划案（GDD）系统架构时使用。在顶层设计
 → 没有变更记录的架构文档，第二轮迭代就会变成黑箱。
 
 ### 2. 系统地图
-Sxx 编号清单（核心系统 2~12 个）+ 支撑层（存档/UI，不拥有核心规则）。
+Sxx 编号清单（核心系统通常 1-5 个，有明确要求可超出 5 个）+ 支撑层（存档/UI，不拥有核心规则）。
 P0 段五列表：
 | 系统 | 目的 | 输入 | 输出 | P0 原因 |
 → 每行 P0 原因必须答"删了它，__ 塌"；答不出的降级或合并。

apps/ai-game-creator-shell/src-tauri/design-agent/resources/skills/concept.md

@@ -69,7 +69,7 @@ description: 写游戏策划案（GDD）概念层时使用。把一句话游戏
 记住三个接口：**对内**锚点仲裁一切；**对下**张力变取舍表、定稿变硬约束；
 **对上**边界画线防止越层。九节不是清单，是一台咬合的机器。
 
-## 四、怎么写（模板即流程，九节按序）
+## 四、怎么写（模板参考结构，建议按此组织）
 （本节是带写法要领的教学版；实际填写的纯净模板在 templates/concept-design.md）
 
 ### 1. 一句话概念

apps/ai-game-creator-shell/src-tauri/design-agent/resources/skills/systems.md

@@ -2,7 +2,7 @@
 
 ---
 name: game-gdd-system-doc
-description: 写单个系统的设计文档（Sxx）时的总纲——通用纪律、十二节同构骨架、
+description: 写单个系统的设计文档（Sxx）时的总纲——通用纪律、十二类常见内容、
   红线与分析文档格式。每类系统的专属写法与模板在 modules/system-types/ 下对应目录的 SKILL.md
   与对应模块的模板.md 里，按需取用。
 ---
@@ -26,9 +26,9 @@ description: 写单个系统的设计文档（Sxx）时的总纲——通用纪
 1. 架构已定稿：找到本系统的 Sxx 编号、职责表行、依赖方向——这是合同。
 2. 在 01~12 文件夹里选最接近的系统类型（可组合，如"钓鱼"=05 采集+06 战斗
    的判定部分），读取对应的 `SKILL.md` 与 `模板.md`。
-3. 该文件夹标注"必读例子"的，先读例子全文做密度锚。
+3. 该文件夹标注"参考例子"的，可先读例子了解写法。
 
-## 三、十二节总览：写什么、为什么、怎么咬合
+## 三、常见内容总览：写什么、为什么、怎么咬合
 
 系统文档回答四个问题：
 **这个系统为什么存在（1~2）→ 玩家怎么用它（3~5）→ 它怎么运转（6~8）→
@@ -52,7 +52,7 @@ description: 写单个系统的设计文档（Sxx）时的总纲——通用纪
 咬合：**对上**服从架构三条合同（编号/职责/依赖）；**对内**状态与接口不越
 职责边界；**对下**第 7 节交接喂 TDD。
 
-## 四、十二节通用写法
+## 四、常见内容的参考写法
 （各系统类型的特殊写法见对应文件夹 SKILL.md；纯净模板在其 模板.md）
 
 1 系统目的：若删除它，__ 会塌——一句话说不出 = 该系统不该存在。
@@ -65,7 +65,7 @@ description: 写单个系统的设计文档（Sxx）时的总纲——通用纪
 8 反馈：每种关键结果给独立反馈形态；失败必须说明原因和恢复路径。
 9 内部循环：动词链；可拆单次/区域/长期三层。
 10 输入输出与依赖：引用具名系统与具名数据，禁泛称"资源"。
-11 边界与非目标：照该类型 skill 的"三不"写全；必含"字段数值归 TDD"一条。
+11 边界与非目标：参考该类型 skill 的“三不”说明边界；建议说明字段与数值的交接边界。
 12 开放问题：结构级才留；手感数值类标"待原型验证"。
 
 ## 五、分析文档（全局一份，按层分节）

apps/ai-game-creator-shell/src-tauri/design-agent/resources/skills/top_design.md

@@ -80,7 +80,7 @@ description: 写游戏策划案（GDD）顶层设计时使用。在概念层定
 （大⇄小⇄最小单位）+ 资源三段全；**对下**系统范围表喂架构的系统地图、
 顶层定稿当架构的紧箍咒、验证标准当原型试玩判据。
 
-## 四、怎么写（模板即流程，十六节按序）
+## 四、怎么写（模板参考结构，建议按此组织）
 （本节是带写法要领的教学版；实际填写的纯净模板在 templates/top-design.md）
 
 ### 1. 顶层定位与规模锚点

apps/ai-game-creator-shell/src-tauri/design-agent/system-prompt.md

@@ -1,12 +1,16 @@
 
 你是游戏策划协作 Agent，与用户持续协作完成游戏设计。像普通策划同事一样交流，使用工作区文件工具读写资料；所有文件路径使用相对路径。根据当前对话、阶段上下文和已有文档决定下一步行动。修改文件后，简要说明修改内容和相对路径。对不确定内容区分用户确认、Agent 建议和待原型验证事项；不要把建议写成用户已确认的决定。
 
-优先完成能够依据已有信息推进的工作，不要为每个设计空白都询问用户。局部、可逆的问题可以先提出合理方案并标为暂定。会影响当前阶段范围、关键规则、下游实现或其他重要方向，且必须由用户决定的问题，应先通过纯文本或问询工具询问，等待用户回答，并据此更新相关产物；不要带着这类未决问题提交阶段审批。
+优先完成能够依据已有信息推进的工作，不要为每个设计空白都询问用户。局部、可逆的问题可以先提出合理方案并标为暂定。会影响当前阶段范围、关键规则、下游实现或其他重要方向，且必须由用户决定的问题，应先通过纯文本或问询工具询问，等待用户回答。决定稳定后，再更新受影响的正式产物和必要的过程记录；不要带着这类未决问题提交阶段审批。
+
+分析阶段优先记录当前目标、上层约束、候选方案、取舍、用户已确认或 Agent 暂定的边界，以及必须检查的验收项。除非用户明确要求展开讨论，不要先在回复中逐节起草与正式文档重复的长篇正文；形成结论后直接写入正式产物，再进行一次必要的一致性检查。文件操作前只需说明简短计划、目标文件和主要变化。
 
 正式策划文档在文档头部写明版本标记，例如“版本：v1”。由你自行维护版本号：只有整体修订、阶段性定稿或用户意见造成实质内容变化时才递增；错别字、措辞润色、单个局部修改和小范围补充不单独递增。
 
 阶段审批是每个阶段的最终检查，表示本阶段产物已经完成，无未决内容，交给用户做最终检阅，不承担问询功能。提交前，解决所有影响本阶段完成的关键问题，或明确说明它们不阻塞本阶段交付，并更新相关产物。可以保留不阻塞当前阶段的后续事项和待原型验证项。
 
+过程文档用于记录关键依据、决定和待办，不要求实时完整，也不应重复正式设计文档。阶段内优先完成主要设计内容；只有稳定且影响后续工作的决定才需要同步到多个过程文档。阶段提交前，补齐影响验收的关键记录。
+
 阶段获批后，产物中已经采用的方案作为后续工作的依据，并保留原有决策来源。除非用户主动质疑或出现新的约束冲突，不要反复要求确认历史暂定决定。
 
 用户说“继续”时，继续推进当前阶段最有价值的工作。判断本阶段已完成并准备交用户检阅时，应调用 `submit_phase_for_approval`；只有该工具调用成功，才算正式提交审批。

apps/ai-game-creator-shell/src-tauri/design-agent/tools.json

@@ -2,7 +2,7 @@
   {"type":"function","function":{"name":"get_workflow_status","description":"读取当前策划工作流状态，只返回阶段列表、当前阶段、已批准阶段和待审批阶段；不推进阶段、不提交审批、不修改文件。","parameters":{"type":"object","properties":{},"additionalProperties":false}}},
   {"type":"function","function":{"name":"list_resources","description":"列出固定资源的逻辑目录、资源 ID、标题和简介。资源是只读的随包文档；不要猜测物理路径。","parameters":{"type":"object","properties":{},"additionalProperties":false}}},
   {"type":"function","function":{"name":"read_resource","description":"读取一份固定资源文档全文。每次读取一个 resource_id；资源只读。读到未实现占位文档时由你自行判断和处理。","parameters":{"type":"object","properties":{"resource_id":{"type":"string"}},"required":["resource_id"],"additionalProperties":false}}},
-  {"type":"function","function":{"name":"patch_file","description":"局部修改 UTF-8 文件，优先用于已有文件的小范围修订。先读文件，以唯一且非空的 old_text 精确匹配并替换为 new_text；new_text 为空可删除片段，保留原文并追加可插入。匹配失败不修改文件。path 使用相对路径。","parameters":{"type":"object","properties":{"path":{"type":"string"},"old_text":{"type":"string"},"new_text":{"type":"string"}},"required":["path","old_text","new_text"],"additionalProperties":false}}},
+  {"type":"function","function":{"name":"patch_file","description":"局部修改 UTF-8 文件。使用 old_text/new_text，或使用 edits 一次进行多个独立替换；每个 old_text 必须非空且在原文件中唯一，匹配失败、重复或范围重叠时不修改文件。path 使用相对路径。","parameters":{"type":"object","properties":{"path":{"type":"string"},"old_text":{"type":"string"},"new_text":{"type":"string"},"edits":{"type":"array","items":{"type":"object","properties":{"old_text":{"type":"string"},"new_text":{"type":"string"}},"required":["old_text","new_text"],"additionalProperties":false}}},"required":["path"],"additionalProperties":false}}},
   {"type":"function","function":{"name":"delete_path","description":"谨慎使用；永久删除工作区内的文件或目录；目录会连同全部内容递归删除，不备份。先确认目标及删除范围。path 使用相对路径，不能删除工作区根目录，也不能经过链接。","parameters":{"type":"object","properties":{"path":{"type":"string"}},"required":["path"],"additionalProperties":false}}},
   {"type":"function","function":{"name":"list_dir","description":"列出工作目录内的文件和目录。path 使用相对路径。","parameters":{"type":"object","properties":{"path":{"type":"string"}},"required":["path"],"additionalProperties":false}}},
   {"type":"function","function":{"name":"read_file","description":"读取工作目录内的 UTF-8 文本文件。path 使用相对路径。","parameters":{"type":"object","properties":{"path":{"type":"string"}},"required":["path"],"additionalProperties":false}}},

apps/ai-game-creator-shell/src-tauri/prompts/runtime/common.md

@@ -1,6 +1,6 @@
 处理代码任务时先用 project.search 定位，再用带行号的 file.read 获取足够上下文；单文件小改优先使用 file.patch；涉及多个文件时优先使用 project.patchset，它会自动创建 checkpoint，无需额外调用 project.checkpoint，并在成功后用返回的 checkpointId 调用 project.diff(includeContent=true) 审查整体变更；只有确认文件已废弃时才删除。
 
-每次成功执行 file.write、file.patch、file.delete、project.patchset 或 project.restore，以及每次真正启动 command.exec 或 command.start，都会产生新的项目 revision；最后一次修改后必须成功执行 project.verify、可验证 command.exec，或成功执行 command.run_limited 的 game.static_smoke，才能调用 respond_to_user 收束。文件回读不能替代可执行验证，验证后再次修改必须重新验证。需要执行 package.json 中的验证脚本时，先读取 package.json，再把真实脚本名和读到的完整命令原样提交给 project.verify；script 可以是 check、typecheck、test、lint、build，或使用 check:<name>、test:<name>（例如 test:unit）、lint:<name>、typecheck:<name>、build:<name>、verify:<name>、validate:<name> 形式的命名脚本，其中冒号后的每个非空段必须以字母或数字开头且只能包含字母、数字、连字符、下划线或点；不得猜测或改写 expectedCommand。
+每次成功执行 file.write、file.patch、file.delete、project.patchset 或 project.restore，以及每次真正启动 command.exec、command.start 或 project.bootstrap，都会产生新的项目 revision；DirectProject 的 npm/Phaser 工程先用 project.bootstrap {cwd:"game"} 执行受控无参数 npm install，再用 project.verify {cwd:"game",script:"build",expectedCommand:从 game/package.json 原样读取} 构建，并确认 game/dist/index.html 存在。最后一次修改后必须成功执行 project.verify、可验证 command.exec，或成功执行 command.run_limited 的 game.static_smoke，才能调用 respond_to_user 收束。文件回读不能替代可执行验证，验证后再次修改必须重新验证。需要执行 package.json 中的验证脚本时，先读取 package.json，再把真实脚本名和读到的完整命令原样提交给 project.verify；script 可以是 check、typecheck、test、lint、build，或使用 check:<name>、test:<name>（例如 test:unit）、lint:<name>、typecheck:<name>、build:<name>、verify:<name>、validate:<name> 形式的命名脚本，其中冒号后的每个非空段必须以字母或数字开头且只能包含字母、数字、连字符、下划线或点；不得猜测或改写 expectedCommand。
 
 每 6 轮只是一次进度 checkpoint 与停滞检测，不是上下文压缩或 run 的终止上限；只要 observation 出现新的独立进展，就在同一 run 继续下一窗口，只有窗口没有新进展时才按停滞处理。真正的上下文压缩仅由 token 阈值或显式 compact 触发。Agent 私有记忆只能由本人写入，跨 Agent 共享稳定结论用 blackboard.write，给单个 Agent 留上下文用 agent.message。
 

apps/ai-game-creator-shell/src-tauri/resources/agc-skills/agc-game-production-workflow/SKILL.md

@@ -0,0 +1,32 @@
+---
+name: agc-game-production-workflow
+description: Orchestrate a complete AGC game delivery from an approved brief to a playable, art-integrated, browser-validated product. Use when creating a new game, implementing a substantial game brief, or turning a planning document into a finished game.
+---
+
+# AGC Game Production Workflow
+
+Use this Skill as the top-level SOP for a new game or a substantial game brief. The tools are stages in one delivery chain, not independent suggestions. Do not stop after producing a plan, after writing code, or after generating an image.
+
+## Stage flow
+
+1. **Brief and scope** — Read the current planning output and project instructions. Extract the game loop, player actions, entities, visual requirements, target viewports, and the completion evidence. If the brief is incomplete, ask focused questions before side effects.
+2. **Project and asset inventory** — Inspect the existing project structure and call `agc_list_registered_assets` (and `agc_list_project_files` when needed). Record which requested visuals already have usable registered identities and which are missing. Do not invent asset identities from filenames.
+3. **Visual production** — For missing or unsuitable visuals, call the reviewed `agc_tools` workflow: use `taonier_prepare_game_art` for a complete package, or `agc_generate_image` / `agc_edit_image` for focused assets. Read returned paths, identities, and warnings. A warning or partial package requires a narrower retry or independent assets before continuing.
+4. **Game implementation** — Implement the complete playable loop and wire the returned project-relative asset paths into the actual runtime. Every required character, object, background, effect, and UI visual must have a real source or an explicit brief-level decision to remain code-native. Generated assets that are unused, documentation-only, or replaced by emoji/CSS placeholders do not satisfy this stage.
+5. **Build and local verification** — Run the project’s bootstrap/install and verify/build commands. Confirm the actual playable entry under `dist` (or the editor runtime for a supported editor project) and fix build or asset-loading failures before preview.
+6. **Browser playtest** — Call `agc_browser_playtest` for desktop and mobile evidence after meaningful changes. Check the game loop, input, layout, asset loading, and visible use of the generated art. Fix findings and repeat stages 4–6 until the evidence is clean.
+7. **Delivery** — Report the implemented behavior, real asset paths and identities used, build result, playtest evidence, warnings, and any explicit remaining gap. Do not claim complete while a required stage is failed, skipped without the brief’s justification, or missing evidence.
+
+## Stage transitions
+
+Advance only when the current stage has its output: brief → inventory; inventory → art decision; art decision → usable registered assets or an explicit no-art decision; implementation → source references to those assets; build → playable entry; playtest → evidence; delivery → truthful report. If a tool fails, preserve its error and handle it under "Error handling" instead of silently substituting a later-stage placeholder.
+
+For a small edit to an existing game where the brief and suitable assets are unchanged, use the focused edit path and do not regenerate art. This exception does not apply to a new game or a substantial planning brief.
+
+## Error handling
+
+When a stage tool, command, or verification fails, retry at most three times before treating that stage as failed. Keep the retries serial and scoped to the same stage and the same input: a retry must not open a parallel path, skip ahead to a later stage, or substitute a placeholder for the missing output.
+
+Only after the third attempt also fails, stop and tell the user the failure reason — which stage failed, which tool or command reported the error, what the error says, and what is still missing. A stage whose three attempts never succeeded is not complete, and its missing output cannot be reported as delivered.
+
+Read the referenced specialist Skills for their detailed contracts: `agc-project-structure`, `taonier-art-assets`, `agc-web-game-development`, `agc-client-projection`, and `agc-browser-playtest`.

apps/ai-game-creator-shell/src-tauri/resources/agc-skills/agc-game-production-workflow/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "完整游戏生产流程"
+  short_description: "从策划案到真实美术接入和试玩验收的连续交付"
+  default_prompt: "Use $agc-game-production-workflow to take the current game brief through inventory, art, implementation, build, playtest, and delivery."

apps/ai-game-creator-shell/src-tauri/resources/agc-skills/agc-game-production-workflow/references/workflow-contract.md

@@ -0,0 +1,5 @@
+# Workflow contract
+
+The production Skill owns sequencing and transition evidence. Specialist Skills own the detailed safety and data rules for each tool family. A specialist tool result is never a delivery result by itself: image generation must be followed by registered identity inspection and runtime integration; code writing must be followed by build verification; a successful preview launch must be followed by desktop and mobile playtest evidence when the brief targets both.
+
+The no-art exception is valid only when the brief explicitly requests a code-native visual treatment or the inventory proves that all required visuals are already registered and suitable. Emoji, CSS primitives, random local files, and generated files that are not referenced by the runtime are not evidence of an integrated art package.

apps/ai-game-creator-shell/src-tauri/resources/agc-skills/agc-web-game-development/SKILL.md

@@ -14,7 +14,7 @@ Implement the user's actual game request in the current project as an npm-manage
 3. Build with the project's npm script before previewing. The playable entry is the package directory's `dist/index.html`; never report an unbuilt bare-module page as playable. Import assets or configure public assets so all runtime media is included in dist; preview and exports cannot read outside it.
 4. Build a complete playable loop: visible objective, responsive input, meaningful state changes, success or failure feedback, and a reliable restart path where the game needs one.
 5. Fit the active game scene to desktop and mobile viewports without accidental page scrollbars. Reserve deliberate safe space for HUD elements instead of covering interactive content.
-6. Reuse registered Taonier art when available through `agc_tools`. Load media defensively and keep gameplay usable when an optional derivative is absent; never relabel a local placeholder as platform art.
+6. Invoke `taonier-art-assets` for every new game brief that needs visual assets. First reuse suitable registered Taonier art; when the brief's required visual elements are missing or unsuitable, call the reviewed `agc_tools` generation/edit workflow in the same task. After the tool returns, wire its relative paths into the game and verify the rendered result. A game with unused generated assets or placeholder emoji/CSS where requested art should appear is not complete. Load media defensively only for genuinely optional effects, and never relabel a local placeholder as platform art.
 7. Let Phaser own the render loop and input dispatch. Avoid duplicate scenes, stale event listeners, and state that survives restart unintentionally.
 8. After a meaningful game change, use the browser playtest Skill and fix issues shown by real evidence before reporting completion.
 

apps/ai-game-creator-shell/src-tauri/resources/agc-skills/manifest.json

@@ -1,7 +1,29 @@
 {
   "schemaVersion": "agc-skill-pack.v1",
-  "version": "2026-08-26.12",
+  "version": "2026-08-26.14",
   "skills": [
+    {
+      "name": "agc-game-production-workflow",
+      "purpose": "把完整游戏从策划案按阶段推进到真实素材接入、构建、试玩和交付",
+      "triggers": [
+        "从策划案创建完整游戏",
+        "实现完整游戏交付",
+        "需要衔接策划、素材、代码、构建和试玩"
+      ],
+      "requiredTools": [
+        "agc_tools.agc_list_registered_assets",
+        "agc_tools.agc_generate_image",
+        "agc_tools.agc_edit_image",
+        "agc_tools.taonier_prepare_game_art",
+        "agc_tools.agc_browser_playtest"
+      ],
+      "files": [
+        "SKILL.md",
+        "agents/openai.yaml",
+        "references/workflow-contract.md"
+      ],
+      "sha256": "d9d8e7e0a6bc512e0b463e0e4bd77edee1cc57f4a6965c9553e0920e38985d5c"
+    },
     {
       "name": "agc-project-structure",
       "purpose": "约束当前项目根、游戏代码、美术素材与客户端状态的职责边界",
@@ -31,6 +53,7 @@
         "已有陶泥儿素材需要接入玩法"
       ],
       "requiredTools": [
+        "agc_tools.agc_list_registered_assets",
         "agc_tools.agc_generate_image",
         "agc_tools.agc_edit_image",
         "agc_tools.taonier_prepare_game_art"
@@ -40,7 +63,7 @@
         "agents/openai.yaml",
         "references/platform-art-contract.md"
       ],
-      "sha256": "82e4b2ee8ca8147b51ca206b0565b3cc244dc5d3cddb8343875001c0beb4711f"
+      "sha256": "bd1e415aac0cd0f97090296f34c67898dd731d1e177ec91a56027f9b68a88b37"
     },
     {
       "name": "agc-web-game-development",
@@ -57,7 +80,7 @@
         "agents/openai.yaml",
         "references/game-quality-checklist.md"
       ],
-      "sha256": "0649c72dd53e05ad7c87b28def1397c2badf61b0c308091196c40f7c48a8b36a"
+      "sha256": "05b5cfbf7a40fd303717491f5cea84ff339a73359c9678b283fd54d2b5c45efd"
     },
     {
       "name": "agc-browser-playtest",