docs: 移除Superpowers文档路径

README.md

@@ -79,7 +79,7 @@ QQBot 插件定时任务由 manifest 的 `tasks` 声明，平台持久化到 `qq
 
 Admin 环境总览面板使用 `ENV_DASHBOARD_*` 只读配置聚合 local-dev、NAS 线上、腾讯云和 r4se 状态。`ENV_DASHBOARD_ADMIN_LOCAL_URL` / `ENV_DASHBOARD_ADMIN_PUBLIC_URL` 只用于展示 Admin 本机与线上入口证据。HTTP 快照提供当前拓扑，后端 local/MQTT 事件总线通过 SSE 推送增量事件给 Admin；前端不直连 MQTT，也不轮询刷新。Jenkins、K8s、Tencent Cloud、Caddy、WireGuard、Mihomo/OpenClash 未配置时会显示 `unwired` 证据，不能渲染成健康假象；第一版不暴露重启、部署、迁移、容器重建、插件启停或代理切换等写操作。
 
-NapCat Runtime/Protocol Profile 已完成本地 API/Admin 实施，线上发布和账号闭环按 `docs/superpowers/plans/2026-06-18-qqbot-napcat-runtime-protocol-profile-implementation-plan.md` 的 Task 10 执行。当前实现覆盖运行态/协议/会话行为/历史登录事件兼容表/风险模式表，真实物理设备风格 hostname/MAC，NapCat/OneBot 配置 hash，KT `zh_CN.UTF-8` 中国桌面派生镜像资产，只读 `/qqbot/napcat/runtime/detail` 证据接口，watchdog 离线巡检告警，以及 Admin 账号页“运行态”抽屉；不绕过 QQ/Tencent 验证码、不修改 QQ/NTQQ 签名协议、不启用 privileged/host network，也不做账号级每小时/每日累计发送预算。NapCat Chinese Desktop Runtime v20 使用 KT `NapCatQQ` fork 源码构建出的 `NapCat.Shell` artifact，并在 QQ `KickedOffLine` 后标记 native login service stale；API 在源 Docker 容器在线但 WebUI 明确 QQ 离线时会同容器调用 `RestartNapCat` 重启 NapCat worker，重建 QQCore login service 后再推进 quick/password/qrcode，不做 Docker 重建、补 env 或设备身份迁移，且同一个更新登录 session 只消费一次 worker restart 预算；v14 起还会对 QQ/NapCat/Xvfb 长期进程的 `/proc/<pid>/mountinfo` 做 PID 级遮蔽，防止 `overlay`、`/vol1/docker`、`docker-init`、`/docker/containers`、`napcat-instances` 等宿主路径泄露；v15 修复扫码成功时 `QQLoginInfo` 晚于登录态写入造成的 QQ 号回读空窗；v16 在 native reset 缺少 `offline()` 时改用 `destroy()` 硬重置半登录服务，并让镜像 verify 等待 mountinfo guard 收敛；v17/v18 增加 WebUI 鉴权的 `/api/Debug/RuntimeViewProbe` 同进程诊断并修正 native maps 截断导致的 hook 证据假阴性；v19 保留 WebUI `RestartNapCat` 重启 worker 时的 `-q <uin>` 快速登录参数，避免重启后退回无账号扫码；v20 保护 API 预写的 `/app/napcat/config`，避免上游首次解包 `NapCat.Shell/*` 覆盖 `bypass.*=true` 与 `o3HookMode=0`。镜像必须先用 `scripts/napcat-desktop-cn-stage-build.mjs` staged build context，生产 `QQBOT_NAPCAT_IMAGE` 应指向验证过的 `kt-napcat-desktop-cn:desktop-cn-v20` digest。`k8s/prod/api.yaml` 保留 `desktop-cn-v20` 稳定默认值；Jenkins `QQBOT_NAPCAT_IMAGE_OVERRIDE` 和 `QQBOT_NAPCAT_DESKTOP_PROFILE_VERSION_OVERRIDE` 仅在填写时通过 `kubectl set env` 推广已验证运行时镜像/profile，空值会继续使用 manifest/default env。回滚时重新运行 Jenkins 并填入上一版 digest/profile，或清空两个 override 后重新部署 manifest 默认值。
+NapCat Runtime/Protocol Profile 已完成本地 API/Admin 实施，线上发布和账号闭环按 `docs/plans/2026-06-18-qqbot-napcat-runtime-protocol-profile-implementation-plan.md` 的 Task 10 执行。当前实现覆盖运行态/协议/会话行为/历史登录事件兼容表/风险模式表，真实物理设备风格 hostname/MAC，NapCat/OneBot 配置 hash，KT `zh_CN.UTF-8` 中国桌面派生镜像资产，只读 `/qqbot/napcat/runtime/detail` 证据接口，watchdog 离线巡检告警，以及 Admin 账号页“运行态”抽屉；不绕过 QQ/Tencent 验证码、不修改 QQ/NTQQ 签名协议、不启用 privileged/host network，也不做账号级每小时/每日累计发送预算。NapCat Chinese Desktop Runtime v20 使用 KT `NapCatQQ` fork 源码构建出的 `NapCat.Shell` artifact，并在 QQ `KickedOffLine` 后标记 native login service stale；API 在源 Docker 容器在线但 WebUI 明确 QQ 离线时会同容器调用 `RestartNapCat` 重启 NapCat worker，重建 QQCore login service 后再推进 quick/password/qrcode，不做 Docker 重建、补 env 或设备身份迁移，且同一个更新登录 session 只消费一次 worker restart 预算；v14 起还会对 QQ/NapCat/Xvfb 长期进程的 `/proc/<pid>/mountinfo` 做 PID 级遮蔽，防止 `overlay`、`/vol1/docker`、`docker-init`、`/docker/containers`、`napcat-instances` 等宿主路径泄露；v15 修复扫码成功时 `QQLoginInfo` 晚于登录态写入造成的 QQ 号回读空窗；v16 在 native reset 缺少 `offline()` 时改用 `destroy()` 硬重置半登录服务，并让镜像 verify 等待 mountinfo guard 收敛；v17/v18 增加 WebUI 鉴权的 `/api/Debug/RuntimeViewProbe` 同进程诊断并修正 native maps 截断导致的 hook 证据假阴性；v19 保留 WebUI `RestartNapCat` 重启 worker 时的 `-q <uin>` 快速登录参数，避免重启后退回无账号扫码；v20 保护 API 预写的 `/app/napcat/config`，避免上游首次解包 `NapCat.Shell/*` 覆盖 `bypass.*=true` 与 `o3HookMode=0`。镜像必须先用 `scripts/napcat-desktop-cn-stage-build.mjs` staged build context，生产 `QQBOT_NAPCAT_IMAGE` 应指向验证过的 `kt-napcat-desktop-cn:desktop-cn-v20` digest。`k8s/prod/api.yaml` 保留 `desktop-cn-v20` 稳定默认值；Jenkins `QQBOT_NAPCAT_IMAGE_OVERRIDE` 和 `QQBOT_NAPCAT_DESKTOP_PROFILE_VERSION_OVERRIDE` 仅在填写时通过 `kubectl set env` 推广已验证运行时镜像/profile，空值会继续使用 manifest/default env。回滚时重新运行 Jenkins 并填入上一版 digest/profile，或清空两个 override 后重新部署 manifest 默认值。
 
 运行时发布时，API 仓库不提交 `NapCat.Shell.zip`；生产镜像必须从 staged context 构建，`fork-artifact.json` 必须带完整 marker metadata，包括 upstream release tag/commit、fork commit、base image digest、Jenkins URL 和 artifact hashes。release evidence 里的 NapCat base image 必须用 digest pin。API Jenkins 只消费人工确认后的运行时推广参数，不自动合并上游、不自动构建隐藏镜像，也不在 override 为空时覆盖 K8s manifest 中的默认 env。
 

docs/plans/2026-06-13-api-runtime-foundation.md

@@ -1,6 +1,6 @@
 # API Runtime Foundation Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Implement Phase 1 of the approved API runtime foundation: typed runtime config, classified runtime errors, structured runtime evidence, and a lightweight `GET /health/runtime` endpoint that can be used by Jenkins/K8s/ktWorkflow in the next plan.
 
@@ -1232,7 +1232,7 @@ pnpm run typecheck
 
 ```powershell
 git status --short
-git add src/runtime src/app.module.ts src/main.ts test/runtime docs/superpowers/plans/2026-06-13-api-runtime-foundation.md
+git add src/runtime src/app.module.ts src/main.ts test/runtime docs/plans/2026-06-13-api-runtime-foundation.md
 git diff --cached --check
 git commit -m "feat: 添加API运行时健康检查"
 ```
@@ -1342,10 +1342,10 @@ git diff --check
 pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --project D:\MyFiles\KT\Node\kt-template-online-api --changed-files README.md API.md src/runtime src/app.module.ts src/main.ts test/runtime
 ```
 
-- [ ] Run the Superpowers code-review gate after implementation:
+- [ ] Run the KT global review gate after implementation:
 
 ```text
-Use superpowers:requesting-code-review, then address actionable findings with superpowers:receiving-code-review if findings are returned.
+Use `KT global review`, then address actionable findings with `KT review follow-up` if findings are returned.
 ```
 
 - [ ] Commit API docs after verification and review pass:
@@ -1391,7 +1391,7 @@ typecheck command and result
 local GET /health/runtime result or explicit local-start blocker
 KT doc sync result
 KT global review result
-Superpowers review result
+KT global review result
 cleanup action for any started local process
 ```
 

docs/plans/2026-06-14-api-deployment-observation-foundation.md

@@ -1,6 +1,6 @@
 # API Deployment Observation Foundation Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Build Phase 2 of the approved API Runtime Foundation: a ktWorkflow deployment observation adapter that gathers Jenkins, K8s, Pod, `/health/runtime`, and task smoke evidence before any deployed feature is called complete.
 
@@ -1458,9 +1458,9 @@ No Critical or Important findings caused by this change.
 Known pre-existing dirty files are classified separately.
 ```
 
-- [ ] **Step 5: Request Superpowers code review**
+- [ ] **Step 5: Request KT global review**
 
-Use `superpowers:requesting-code-review` after implementation and before declaring the work complete. Review must cover:
+Use `KT global review` after implementation and before declaring the work complete. Review must cover:
 
 ```text
 mcp/ktWorkflow deploy observation command safety
@@ -1480,7 +1480,7 @@ All Critical and Important findings are fixed or explicitly rejected with eviden
 Run:
 
 ```powershell
-pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run workstream-closeout -- --title "API Runtime Foundation Phase 2 部署观测" --project mcp --project api --verification "typecheck/self-test/deploy-observation dry-run/read-only observation evidence" --doc-sync "kt_change_doc_sync README handled; API health contract unchanged" --cleanup "cleanup-history final dry-run deleted=[]" --cleanup-final-deleted 0 --review "global-review no Critical/Important findings caused by this change" --superpowers-review "Superpowers reviewer completed; no Critical/Important findings remain" --problem "无新卡点或已记录 NAS/Jenkins/K8s 只读观测阻塞" --solution "kt_deploy_observation 固化 Jenkins/K8s/health/smoke 观测" --pattern deploy-observation --kt-workflow-updated --upgrade-notes "新增 kt_deploy_observation 工具和 CLI"
+pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run workstream-closeout -- --title "API Runtime Foundation Phase 2 部署观测" --project mcp --project api --verification "typecheck/self-test/deploy-observation dry-run/read-only observation evidence" --doc-sync "kt_change_doc_sync README handled; API health contract unchanged" --cleanup "cleanup-history final dry-run deleted=[]" --cleanup-final-deleted 0 --review "global-review no Critical/Important findings caused by this change" --problem "无新卡点或已记录 NAS/Jenkins/K8s 只读观测阻塞" --solution "kt_deploy_observation 固化 Jenkins/K8s/health/smoke 观测" --pattern deploy-observation --kt-workflow-updated --upgrade-notes "新增 kt_deploy_observation 工具和 CLI"
 ```
 
 Expected:

docs/plans/2026-06-15-api-admin-architecture-convergence-v3.md

@@ -1,12 +1,12 @@
 # API/Admin Architecture Convergence V3 Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Make the third-phase target architecture true in code by migrating all API legacy roots into `src/modules/**`, enforcing structure tests, slimming every API/Admin domain, preserving behavior, and deleting `src/admin`, `src/blog`, `src/minio`, `src/wordpress`, and `src/qqbot`.
 
 **Architecture:** Work from hard RED structure gates, then migrate one domain at a time. Each domain task includes three inseparable pieces: move ownership into the target module, remove duplicate/unused transitional code, and run behavior checks that prove existing routes and Admin pages still work. The final task deletes old roots and runs the full convergence gate.
 
-**Tech Stack:** NestJS 11, TypeORM 0.3, MySQL, Jest, pnpm 9 API workspace, Vben Admin 5, Vue 3 TSX/Vue SFC, Vitest, pnpm 10 Admin workspace, ktWorkflow global review, Superpowers review.
+**Tech Stack:** NestJS 11, TypeORM 0.3, MySQL, Jest, pnpm 9 API workspace, Vben Admin 5, Vue 3 TSX/Vue SFC, Vitest, pnpm 10 Admin workspace, ktWorkflow global review, KT global review.
 
 ---
 
@@ -1227,7 +1227,7 @@ Create `docs/refactor-v3/architecture-convergence-verification.md`:
 ## Review
 
 - KT global review: pass.
-- Superpowers review: pass.
+- KT global review: pass.
 ```
 
 - [ ] **Step 8: Update TASKS**
@@ -1239,7 +1239,7 @@ Update `D:\MyFiles\KT\TASKS.md` with a short record:
 
 - 范围：`Node/kt-template-online-api`、`Vue/kt-template-admin`、`TASKS.md`。
 - 关键词：旧根删除、`src/modules/**` 强门禁、全模块瘦身、Admin 全域边界、结构测试、API/Admin 本地验证。
-- 验证：记录结构 gate、API typecheck/Jest、Admin Vitest/typecheck、local smoke、global review、Superpowers review 的最终结果。
+- 验证：记录结构 gate、API typecheck/Jest、Admin Vitest/typecheck、local smoke、KT global review 的最终结果。
 ```
 
 - [ ] **Step 9: Run review gates**
@@ -1252,7 +1252,7 @@ pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --contentScanMode c
 
 Expected: `findings=[]`.
 
-Use Superpowers requesting-code-review before declaring implementation complete.
+Use KT workflow KT global review before declaring implementation complete.
 
 - [ ] **Step 10: Commit final verification**
 
@@ -1313,7 +1313,7 @@ global-review findings=[].
 | Behavior preservation | Focused Jest, typecheck, local smoke, Admin Vitest/typecheck |
 | No meaningless backup artifacts | Execution Rules and commit scopes |
 | Documentation and TASKS update | Tasks 0 and 7 |
-| KT global review and Superpowers review | Task 7 |
+| KT global review | Task 7 |
 
 ### Placeholder Scan
 

docs/plans/2026-06-15-api-admin-full-refactor-v3.md

@@ -1,10 +1,10 @@
 # API/Admin Full Refactor V3 Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Complete the third-phase API/Admin full-module refactor from the approved design: rebuild schema, migrate all API modules, rebuild the QQBot plugin platform, finish NapCat device/login flow, sync Admin, and close locally before one unified push and online smoke.
 
-**Architecture:** Execute as a batch-gated migration. API owns schema, domain contracts, runtime state, plugin execution, and deployment smoke; Admin follows API contracts in the same batch or the next immediate task. Each batch starts with RED checks, ends with local verification, KT review, Superpowers review, and separate API/Admin commits.
+**Architecture:** Execute as a batch-gated migration. API owns schema, domain contracts, runtime state, plugin execution, and deployment smoke; Admin follows API contracts in the same batch or the next immediate task. Each batch starts with RED checks, ends with local verification, KT review, KT global review, and separate API/Admin commits.
 
 **Tech Stack:** NestJS 11, TypeORM 0.3, MySQL, pnpm 9 API workspace, Vben Admin 5 / Vue 3 / TSX / Ant Design Vue, pnpm 10 Admin workspace, Jest, Playwright, ktWorkflow MCP, Jenkins/K8s/NapCat runtime.
 
@@ -1534,10 +1534,10 @@ kt_change_doc_sync(...)
 kt_global_code_review(...)
 ```
 
-Run Superpowers review before claiming a batch is finished:
+Run KT global review before claiming a batch is finished:
 
 ```text
-superpowers:requesting-code-review
+`KT global review`
 ```
 
 If review finds a real Important issue, fix it and re-review before commit.
@@ -1584,5 +1584,5 @@ The third phase is complete only when all of these are true:
 - Online functional smoke passes.
 - NapCat real account flow completes, including device persistence and new-device flow when triggered.
 - TASKS and required docs are updated.
-- KT global review and Superpowers review are clean.
+- KT global review are clean.
 - No Node/Vite validation process started by the run remains alive.

docs/plans/2026-06-15-qqbot-architecture-convergence-bangdream-rewrite.md

@@ -1,6 +1,6 @@
 # QQBot Architecture Convergence And BangDream Rewrite Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Make the approved third-phase QQBot target architecture true in code: core, plugin-platform, napcat, built-in plugins, BangDream, Admin callers/pages, local validation, deployment observation, and online account smoke must all close.
 
@@ -22,7 +22,7 @@ This plan is complete only when all of these are true:
 - FF14 Market, FFLogs, and Repeater are rebuilt as `ff14-market`, `fflogs`, and `repeater` packages with no Nest/Admin/Core/env/direct HTTP or direct fs dependencies.
 - NapCat device identity, container runtime, login session, challenge, new-device flow, and cleanup failure semantics are recoverable from persistence.
 - Admin callers and pages are split by QQBot core, Plugin Platform, and NapCat state domains.
-- Local API/Admin tests, local real interface smoke, browser smoke, KT global review, Superpowers review, push/deploy observation, and online QQBot/NapCat smoke are recorded.
+- Local API/Admin tests, local real interface smoke, browser smoke, KT global review, KT global review, push/deploy observation, and online QQBot/NapCat smoke are recorded.
 
 ## Current Baseline
 
@@ -816,7 +816,7 @@ pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --project admin --c
 pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --project root --content-scan-mode changed
 ```
 
-- [ ] Use Superpowers `requesting-code-review` before claiming local closure.
+- [ ] Use KT workflow `KT global review` before claiming local closure.
 - [ ] Resolve real Important findings and re-run focused review.
 
 ## Batch 10: Push, Deploy Observation, And Online Closure
@@ -907,8 +907,8 @@ pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --project admin --c
 pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --project root --content-scan-mode changed
 ```
 
-- [ ] Run Superpowers `verification-before-completion`.
+- [ ] Run KT workflow `KT completion audit`.
-- [ ] Run Superpowers `finishing-a-development-branch`.
+- [ ] Run KT workflow `KT branch closeout`.
 - [ ] Report final evidence with:
   - API/Admin/root commit hashes
   - pushed branch names
@@ -922,7 +922,7 @@ pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --project root --co
 - Every batch that changes code runs `git diff --check`.
 - Every batch that changes API/Admin contracts updates `API.md`, `docs/refactor-v3/**`, SQL seed/schema, Admin caller, and tests in the same or immediately following batch.
 - Every batch commit is preceded by KT global review for the changed repo.
-- After implementation batches, use Superpowers `requesting-code-review` before claiming the batch done.
+- After implementation batches, use KT workflow `KT global review` before claiming the batch done.
 - If review reports a real Important issue, fix and re-run the relevant review before commit.
 - If a reviewer reports a false positive caused by workflow rules, solidify the false-positive handling in `mcp/ktWorkflow` instead of leaving it as chat context.
 
@@ -946,4 +946,4 @@ pnpm --dir D:\MyFiles\KT\mcp\ktWorkflow run global-review -- --project root --co
 
 ## Implementation Entry
 
-Start execution with Superpowers `subagent-driven-development`. Use subagents only for independent read-only audit or code-review work with a Context Packet. Main-thread implementation should begin at Batch 0 and must not skip RED gates in Batch 1.
+Start execution with KT workflow `KT batch execution`. Use subagents only for independent read-only audit or code-review work with a Context Packet. Main-thread implementation should begin at Batch 0 and must not skip RED gates in Batch 1.

docs/plans/2026-06-16-qqbot-plugin-scheduled-tasks-implementation-plan.md

@@ -1,6 +1,6 @@
 # QQBot Plugin Scheduled Tasks Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Build a unified QQBot plugin scheduled-task platform, expose it in Admin, and land BangDream Bestdori main-data sync as the first managed plugin task.
 
@@ -14,7 +14,7 @@
 
 - API repo: `D:\MyFiles\KT\Node\kt-template-online-api`
 - Admin repo: `D:\MyFiles\KT\Vue\kt-template-admin`
-- Design spec: `D:\MyFiles\KT\Node\kt-template-online-api\docs\superpowers\specs\2026-06-16-qqbot-plugin-scheduled-tasks-design.md`
+- Design spec: `D:\MyFiles\KT\Node\kt-template-online-api\docs\specs\2026-06-16-qqbot-plugin-scheduled-tasks-design.md`
 - Online interface stage is already closed: final smoke artifact `.kt-workspace/test-artifacts/online-full-smoke/20260616-164531/api-full-interface-smoke-20260616091328.json` reported 175 OpenAPI operations, 102 pass, 4 expectedBlocked, 69 skipped, 0 failed. QQBot concurrent `/查分数表 cn` + `/查卡 472` both passed after the queue-wait timeout fix deployed as API `14e2ec9`.
 - Existing plugin-platform owns manifest parsing, worker runtime, persistence entities, controller, and operation/event capabilities under `src/modules/qqbot/plugin-platform/**`.
 - Existing BangDream package root is `src/modules/qqbot/plugins/bangdream`; do not create transfer layers or `builtins/**`.

docs/plans/2026-06-18-admin-environment-dashboard-implementation-plan.md

@@ -1,6 +1,6 @@
 # Admin Environment Dashboard Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** 将 Admin `/dashboard/analytics` 改造成多站点环境状态总览总控面板，覆盖本机开发、NAS 线上、腾讯云、r4se 远程环境，并明确展示未接入证据，第一版只做观测与只读自检，不提供高风险写操作。
 
@@ -12,7 +12,7 @@
 
 ## Source Spec
 
-- `D:\MyFiles\KT\Node\kt-template-online-api\docs\superpowers\specs\2026-06-18-admin-environment-dashboard-design.md`
+- `D:\MyFiles\KT\Node\kt-template-online-api\docs\specs\2026-06-18-admin-environment-dashboard-design.md`
 - 当前确认的页面方向：方案 A `Site Command Center`
 - 第一版能力边界：观测与只读自检；重启、部署、迁移、重建容器、插件启停、任务立即执行、Caddy/OpenClash 改配置等动作必须展示为禁用。
 
@@ -755,7 +755,7 @@ const severityWeight: Record<EnvironmentHealthStatus, number> = {
   ```powershell
   cd D:\MyFiles\KT\Node\kt-template-online-api
   git status --short
-  git add package.json pnpm-lock.yaml .env.example README.md API.md src/modules/admin/platform-config src/modules/qqbot/core/qqbot-core.module.ts src/modules/qqbot/plugin-platform/qqbot-plugin-platform.module.ts test/modules/admin/environment-dashboard docs/superpowers/plans/2026-06-18-admin-environment-dashboard-implementation-plan.md
+  git add package.json pnpm-lock.yaml .env.example README.md API.md src/modules/admin/platform-config src/modules/qqbot/core/qqbot-core.module.ts src/modules/qqbot/plugin-platform/qqbot-plugin-platform.module.ts test/modules/admin/environment-dashboard docs/plans/2026-06-18-admin-environment-dashboard-implementation-plan.md
   git commit -m "feat: 新增环境总览总控面板接口"
   ```
   Expected output: one API commit. Do not stage secrets.

docs/plans/2026-06-18-qqbot-napcat-device-name-timezone-hardening-plan.md

@@ -1,6 +1,6 @@
 # QQBot NapCat Device Name And Timezone Hardening Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Restore a visible QQ security-page device name while keeping NapCat device identity persistent, fix mixed UTC/Asia-Shanghai timestamps, and make runtime/profile evidence persist online.
 

docs/plans/2026-06-18-qqbot-napcat-runtime-protocol-profile-implementation-plan.md

@@ -1,6 +1,6 @@
 # QQBot NapCat Runtime Protocol Profile Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Implement the confirmed NapCat session behavior, login-event minimization, protocol risk profile, real-device identity migration, and Chinese Desktop Runtime profile without adding unsafe automation or hourly/daily send budgets.
 
@@ -16,7 +16,7 @@
   - API: `D:/MyFiles/KT/Node/kt-template-online-api`
   - Admin: `D:/MyFiles/KT/Vue/kt-template-admin`
   - Root docs index: `D:/MyFiles/KT/TASKS.md` after files change.
-- Source spec: `D:/MyFiles/KT/Node/kt-template-online-api/docs/superpowers/specs/2026-06-18-qqbot-napcat-linux-runtime-protocol-profile-design.md`
+- Source spec: `D:/MyFiles/KT/Node/kt-template-online-api/docs/specs/2026-06-18-qqbot-napcat-linux-runtime-protocol-profile-design.md`
 - Preserve behavior:
   - Existing scan create/refresh, captcha, new-device verification, OneBot reverse WS, and QQBot command execution must keep working.
   - Existing watchdog order remains quick login, then saved password login; it must not enter QR scanning automatically.
@@ -2051,7 +2051,7 @@ git -C D:/MyFiles/KT/Vue/kt-template-admin commit -m "feat: 展示NapCat运行
 **Files:**
 - Modify: `API.md`
 - Modify: `README.md`
-- Modify: `docs/superpowers/specs/2026-06-18-qqbot-napcat-linux-runtime-protocol-profile-design.md` only if implementation changes the confirmed design boundary.
+- Modify: `docs/specs/2026-06-18-qqbot-napcat-linux-runtime-protocol-profile-design.md` only if implementation changes the confirmed design boundary.
 - Modify: `D:/MyFiles/KT/docs/qqbot-nas-runtime.md`
 - Modify: `D:/MyFiles/KT/docs/obsidian/modules/KT 模块 - QQBot BangDream FFLogs.md`
 - Modify: `D:/MyFiles/KT/TASKS.md`

docs/plans/2026-06-18-qqbot-napcat-runtime-protocol-profile-implementation-plan.zh-CN.md

@@ -8,8 +8,8 @@
 
 ## 当前状态
 
-- 设计文档已确认：`docs/superpowers/specs/2026-06-18-qqbot-napcat-linux-runtime-protocol-profile-design.md`。
+- 设计文档已确认：`docs/specs/2026-06-18-qqbot-napcat-linux-runtime-protocol-profile-design.md`。
-- 英文实施计划已生成：`docs/superpowers/plans/2026-06-18-qqbot-napcat-runtime-protocol-profile-implementation-plan.md`。
+- 英文实施计划已生成：`docs/plans/2026-06-18-qqbot-napcat-runtime-protocol-profile-implementation-plan.md`。
 - 当前还没有进入代码实现阶段，线上规则仍按现有 NapCat 链路运行。
 - 用户已确认：现有账号已经进入风控，可以做受控设备身份迁移；MAC 不走 Docker/QEMU/KVM/VMware/Hyper-V 风格前缀；`zh_CN.UTF-8` 派生镜像和真实中国桌面环境必须做；不做账号级每小时/每日发送预算。
 
@@ -309,5 +309,5 @@ Admin 首版做只读 Drawer：
 
 - 本中文文档用于确认方向、边界、风险和验收。
 - 英文实施计划用于执行，里面有具体文件、测试、提交节奏和任务拆分。
-- 后续开始实现时，应按英文计划走 `superpowers:subagent-driven-development` 或 `superpowers:executing-plans`。
+- 后续开始实现时，应按英文计划走 `KT batch execution` 或 `KT local execution`。
 - 如果实现阶段发现英文计划与本中文方案冲突，以本中文方案的目标和边界为准，再同步修订英文计划。

docs/plans/2026-06-18-qqbot-plugin-platform-decoupling-implementation-plan.md

@@ -1,6 +1,6 @@
 # QQBot Plugin Platform Decoupling Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Make `src/modules/qqbot/plugin-platform` a generic plugin runtime that never imports or branches on BangDream, FF14 Market, FFLogs, or Repeater package code.
 
@@ -1818,7 +1818,7 @@ $patterns = @(
   'write tests' + ' for'
 )
 foreach ($pattern in $patterns) {
-  rg -n --fixed-strings $pattern docs/superpowers/plans/2026-06-18-qqbot-plugin-platform-decoupling-implementation-plan.md
+  rg -n --fixed-strings $pattern docs/plans/2026-06-18-qqbot-plugin-platform-decoupling-implementation-plan.md
 }
 ```
 

docs/plans/2026-06-19-qqbot-bilibili-card-plugin-implementation-plan.md

@@ -1,6 +1,6 @@
 # QQBot Bilibili Card Plugin Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Build a QQBot built-in event plugin that parses Bilibili links from QQ/NapCat card messages and replies with a concise video summary.
 

docs/plans/2026-06-23-qqbot-napcat-source-fork-qr-refresh-implementation-plan.md

@@ -1,6 +1,6 @@
 # QQBot NapCat Source Fork QR Refresh Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Build and ship a source-level NapCatQQ fork fix so stale WebUI login state no longer blocks fresh QR generation, then consume that fork through the KT `desktop-cn-v3` NapCat image.
 

docs/plans/2026-06-23-qqbot-napcat-source-fork-qr-refresh-implementation-plan.zh-CN.md

@@ -1,6 +1,6 @@
 # QQBot NapCat 源码 Fork 二维码刷新修复实施计划（中文）
 
-> **给执行代理的要求：** 实施时必须使用 `superpowers:subagent-driven-development`（推荐）或 `superpowers:executing-plans`，按任务逐项执行并用复选框追踪。英文计划 `2026-06-23-qqbot-napcat-source-fork-qr-refresh-implementation-plan.md` 是精确执行版，包含完整测试代码和源码片段；本文件是同任务、同顺序、同验收标准的中文审核版，供用户确认范围和节奏。
+> **执行说明：** 按 KT 本地工作流逐项执行，并用文内复选框跟踪计划状态。
 
 **目标：** fork NapCatQQ 源码，在源代码层修复 WebUI 登录态 stale 导致二维码无法刷新的问题，并通过 KT `desktop-cn-v3` 派生镜像上线验证。
 

docs/plans/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-implementation-plan.md

@@ -1,6 +1,6 @@
 # NapCatQQ Upstream Sync Runtime Release Pipeline Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Build a ktWorkflow-owned automation path that audits upstream NapCatQQ releases, assists review with Codex CLI, builds verified KT NapCat runtime images, and promotes them to API through explicit deployment parameters.
 
@@ -1026,11 +1026,11 @@ git -C D:\MyFiles\KT commit -m "docs: 记录NapCatQQ自动化实施结果"
 
 ## Execution Handoff
 
-Plan complete and saved to `docs/superpowers/plans/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-implementation-plan.md`.
+Plan complete and saved to `docs/plans/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-implementation-plan.md`.
 
 Two execution options:
 
 1. **Subagent-Driven (recommended)** - Dispatch a fresh subagent per task, review between tasks, and keep ktWorkflow/API/NapCatQQ commits independent.
-2. **Inline Execution** - Execute tasks in this session using `superpowers:executing-plans`, with review checkpoints after each repository boundary.
+2. **Inline Execution** - Execute tasks in this session using `KT local execution`, with review checkpoints after each repository boundary.
 
 Recommended choice: **1**. This work touches three repositories and two runtime environments, so task-level subagents plus parent review will keep scope cleaner.

docs/plans/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-implementation-plan.zh-CN.md

@@ -1,6 +1,6 @@
 # NapCatQQ 上游同步与运行时发布流水线实施计划
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **执行说明：** 按 KT 本地工作流逐项执行，并用文内复选框跟踪计划状态。
 
 **Goal:** 把 NapCatQQ 上游 release 审计、Codex CLI 辅助审查、KT 派生运行时镜像构建、API 显式参数推广全部收敛到可验证的 ktWorkflow 自动化链路。
 
@@ -608,11 +608,11 @@ git -C D:\MyFiles\KT commit -m "docs: 记录NapCatQQ自动化实施结果"
 
 ## 执行交接
 
-计划已保存到 `docs/superpowers/plans/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-implementation-plan.zh-CN.md`。
+计划已保存到 `docs/plans/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-implementation-plan.zh-CN.md`。
 
 两种执行方式：
 
 1. **Subagent-Driven（推荐）**：每个 task 派一个新 subagent，父线程逐 task review，ktWorkflow/API/NapCatQQ 分仓提交。
-2. **Inline Execution**：本线程用 `superpowers:executing-plans` 批次执行，每个仓库边界做 review checkpoint。
+2. **Inline Execution**：本线程用 `KT local execution` 批次执行，每个仓库边界做 review checkpoint。
 
 推荐选 **1**。这条线跨三个仓库和两个运行环境，用 task 级 subagent 更容易守住边界。

docs/plans/2026-06-24-qqbot-napcat-webui-gateway-implementation-plan.md

@@ -1,6 +1,6 @@
 # QQBot NapCat WebUI Gateway Implementation Plan
 
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+> **Execution note:** Execute this plan task-by-task with the KT-local workflow and use the checkboxes to track plan state.
 
 **Goal:** Build an independent NapCat WebUI Gateway service and an Admin second-level route that opens the selected QQBot account's original NapCat WebUI with full operation capability and route-bound session cleanup.
 
@@ -12,8 +12,8 @@
 
 ## Source References
 
-- Spec: `docs/superpowers/specs/2026-06-24-qqbot-napcat-webui-gateway-design.md`
+- Spec: `docs/specs/2026-06-24-qqbot-napcat-webui-gateway-design.md`
-- Chinese spec: `docs/superpowers/specs/2026-06-24-qqbot-napcat-webui-gateway-design.zh-CN.md`
+- Chinese spec: `docs/specs/2026-06-24-qqbot-napcat-webui-gateway-design.zh-CN.md`
 - `http-proxy-middleware` supports Express proxy middleware and WebSocket upgrades through `ws: true` and upgrade handling: <https://github.com/chimurai/http-proxy-middleware>
 - `http-proxy-middleware` WebSocket recipe documents `ws: true`, manual `server.on('upgrade', proxy.upgrade)`, multiple targets, and path rewriting: <https://github.com/chimurai/http-proxy-middleware/blob/master/recipes/websocket.md>
 - `@nestjs-modules/ioredis` provides Nest `RedisModule.forRoot` and `@InjectRedis()` over `ioredis`: <https://github.com/nest-modules/ioredis>
@@ -1864,11 +1864,11 @@ Report:
 
 ## Execution Handoff
 
-Plan complete and saved to `docs/superpowers/plans/2026-06-24-qqbot-napcat-webui-gateway-implementation-plan.md`.
+Plan complete and saved to `docs/plans/2026-06-24-qqbot-napcat-webui-gateway-implementation-plan.md`.
 
 Two execution options:
 
 1. **Subagent-Driven (recommended)** - Dispatch a fresh subagent per task, review between tasks, fast iteration.
-2. **Inline Execution** - Execute tasks in this session using executing-plans, batch execution with checkpoints.
+2. **Inline Execution** - Execute tasks in this session using the KT-local execution loop, batch execution with checkpoints.
 
 Which approach?

docs/plans/2026-06-24-qqbot-napcat-webui-gateway-implementation-plan.zh-CN.md

@@ -1,6 +1,6 @@
 # QQBot NapCat WebUI Gateway 实施计划
 
-> **给 agent worker：** 必须使用 `superpowers:subagent-driven-development`（推荐）或 `superpowers:executing-plans` 按任务执行本计划。步骤使用 checkbox（`- [ ]`）跟踪。
+> **执行说明：** 按 KT 本地工作流逐项执行，并用文内复选框跟踪计划状态。
 
 **目标：** 实现独立 NapCat WebUI Gateway 微服务，并在 Admin 二级页面中打开指定 QQBot 账号的原版 NapCat WebUI，支持完整操作和页面生命周期清理。
 
@@ -12,8 +12,8 @@
 
 ## 来源参考
 
-- 设计文档：`docs/superpowers/specs/2026-06-24-qqbot-napcat-webui-gateway-design.md`
+- 设计文档：`docs/specs/2026-06-24-qqbot-napcat-webui-gateway-design.md`
-- 中文设计：`docs/superpowers/specs/2026-06-24-qqbot-napcat-webui-gateway-design.zh-CN.md`
+- 中文设计：`docs/specs/2026-06-24-qqbot-napcat-webui-gateway-design.zh-CN.md`
 - `http-proxy-middleware` 支持 Express proxy middleware 和 WebSocket upgrade：<https://github.com/chimurai/http-proxy-middleware>
 - `http-proxy-middleware` WebSocket recipe 说明了 `ws: true`、手动 `server.on('upgrade', proxy.upgrade)`、多 target 和 path rewrite：<https://github.com/chimurai/http-proxy-middleware/blob/master/recipes/websocket.md>
 - `@nestjs-modules/ioredis` 提供 Nest `RedisModule.forRoot` 和 `@InjectRedis()`，底层使用 `ioredis`：<https://github.com/nest-modules/ioredis>
@@ -871,11 +871,11 @@ https://admin.kwitsukasa.top/#/qqbot/account
 
 ## 执行交接
 
-计划已保存到 `docs/superpowers/plans/2026-06-24-qqbot-napcat-webui-gateway-implementation-plan.zh-CN.md`。
+计划已保存到 `docs/plans/2026-06-24-qqbot-napcat-webui-gateway-implementation-plan.zh-CN.md`。
 
 两个执行选项：
 
 1. **Subagent-Driven（推荐）**：每个任务派一个新 subagent，任务之间主线程 review，迭代更快。
-2. **Inline Execution**：在当前会话用 executing-plans 执行，按批次检查。
+2. **Inline Execution**：在当前会话用 KT 本地执行流程 执行，按批次检查。
 
 你选哪个？

docs/specs/2026-06-13-api-runtime-foundation-design.md

@@ -67,7 +67,7 @@ This spec covers the first API-wide subproject:
   module boundaries, contract stability, and testability without large,
   unrelated rewrites.
 
-Future subprojects should receive their own Superpowers specs:
+Future subprojects should receive their own KT specifications:
 
 - API module boundary refactor.
 - API contract and DTO stability.
@@ -467,7 +467,7 @@ Completion of the full workstream requires:
 - KT documentation sync.
 - cleanup-history final `deleted=[]` evidence when artifacts are produced.
 - KT global review with no blocking findings.
-- Superpowers code review evidence.
+- KT global review evidence.
 - ktWorkflow closeout evidence.
 
 ## Worktree Cleanliness

docs/specs/2026-06-14-api-full-module-architecture-schema-plugin-platform-design.md

@@ -25,8 +25,8 @@ Admin、Blog、WordPress、MinIO、QQBot、NapCat、Loki 运行期事件和线
 4. 定义后续 C 阶段的迁移矩阵和实施顺序。
 
 本文件不实现代码、SQL、Admin 页面或运行时行为。用户 review 通过后，下一步
-进入 Superpowers `writing-plans`。完整 C 阶段迁移开始前，必须重新走一轮
+进入 KT workflow `KT plan writing`。完整 C 阶段迁移开始前，必须重新走一轮
-Superpowers brainstorming。
+KT requirements and design review。
 
 ## 目标
 
@@ -390,15 +390,15 @@ SSE 事件名、Admin API wrapper 或 SQL seed 标识，都必须在编码前写
 
 本阶段完成条件：
 
-1. 设计文档写入 `docs/superpowers/specs/`。
+1. 设计文档写入 `docs/specs/`。
 2. 设计文档已提交。
 3. 用户 review 已提交文档。
-4. 下一步进入 Superpowers `writing-plans`，不是直接实现代码。
+4. 下一步进入 KT workflow `KT plan writing`，不是直接实现代码。
-5. 后续完整迁移开始前，重新走 Superpowers brainstorming。
+5. 后续完整迁移开始前，重新走 KT requirements and design review。
 
 ## 交接到下一阶段
 
-用户 review 通过后，`writing-plans` 阶段应产出计划文档，把本文转换成可执行 work packages、
+用户 review 通过后，`KT plan writing` 阶段应产出计划文档，把本文转换成可执行 work packages、
 schema 文档、插件平台任务和验证清单。
 
 后续完整实现必须重新 brainstorming，因为它会涉及破坏式 schema 工作、Admin/API 契约决策、

docs/specs/2026-06-15-api-admin-architecture-convergence-v3-design.md

@@ -101,7 +101,7 @@ Admin 当前结构证据：
 - API 行为验证：typecheck、聚焦 Jest、核心接口真实本地 smoke。
 - Admin 行为验证：typecheck、关键 QQBot/插件/NapCat 页面 smoke。
 - 文档同步：架构收敛说明、必要 API/Admin 契约说明、`TASKS.md` 简短记录。
-- KT global review 与 Superpowers code review。
+- KT global review。
 
 ## 非目标
 
@@ -315,7 +315,7 @@ Admin 瘦身完成证据：
 - API 本地真实 smoke 覆盖 `/health/runtime`、Admin 登录/菜单、Blog public、Asset 上传/下载、QQBot command test、插件平台、NapCat 模拟登录状态机。
 - Admin typecheck 通过。
 - Admin 关键页面 smoke 通过。
-- KT global review 与 Superpowers review 无阻断。
+- KT global review 无阻断。
 
 ## 提交策略
 
@@ -351,9 +351,9 @@ Admin 瘦身完成证据：
 - 代码 review 门禁通过。
 - 工作区最终干净。
 
-## 交接到 writing-plans
+## 交接到 KT plan writing
 
-用户审阅并确认本 spec 后，下一步进入 Superpowers `writing-plans`。计划必须把本 spec 拆成可执行任务，尤其要写清：
+用户审阅并确认本 spec 后，下一步进入 KT workflow `KT plan writing`。计划必须把本 spec 拆成可执行任务，尤其要写清：
 
 - 结构测试的 RED/GREEN 步骤和精确文件路径。
 - 每个旧根迁移到哪个目标模块。

docs/specs/2026-06-15-api-admin-full-refactor-v3-design.md

@@ -6,7 +6,7 @@
 
 当前业务数据不重要，因此第三期允许从头设计数据库表，允许破坏式重建 schema，允许重写旧初始化 SQL 和旧模块结构。破坏式不等于无保护：本阶段必须把备份、回滚、验证 SQL、本地 dry run、线上发布观测和真实功能 smoke 作为同一条闭环的一部分。
 
-本文件是 Superpowers brainstorming 的第三期设计落地文档。它只固化设计和执行边界，不创建实现分支，不清理 Admin 旧产物，不写业务实现。用户 review 本文件后，下一步进入 Superpowers `writing-plans`。
+本文件是 KT requirements and design review 的第三期设计落地文档。它只固化设计和执行边界，不创建实现分支，不清理 Admin 旧产物，不写业务实现。用户 review 本文件后，下一步进入 KT workflow `KT plan writing`。
 
 ## 已确认决策
 
@@ -35,7 +35,7 @@
 - 现有 QQBot 插件重写。
 - NapCat Runtime。
 - 全量初始化 SQL、schema 文档、验证 SQL、备份和恢复脚本。
-- 本地、发布、线上 smoke 和 KT/Superpowers review 证据。
+- 本地、发布、线上 smoke 和 KT global review 证据。
 
 ### Admin 仓库
 
@@ -320,7 +320,7 @@ SSE/Admin 状态至少覆盖：
 
 - 空库 schema dry run 可执行。
 - 初始 seed 覆盖 Admin 登录、菜单、基础设置、插件 metadata 和 QQBot 基础命令。
-- writing-plans 输出的任务能够映射到 Batch 1-8。
+- KT plan writing 输出的任务能够映射到 Batch 1-8。
 
 ### Batch 1：Runtime/Common
 
@@ -465,7 +465,7 @@ SSE/Admin 状态至少覆盖：
 3. 本地验证：按批次风险运行 typecheck、scoped lint、scoped Jest、真实 API 请求和 Admin smoke。
 4. 文档同步：更新 schema、README/API、TASKS 或相关 docs。
 5. KT global review。
-6. Superpowers code review。
+6. KT global review。
 7. 批次提交，API/Admin 分仓提交。
 
 接口变化必须真实本地调用一次。发布或线上功能必须通过对应线上 self-test 后才能声明完成。
@@ -498,16 +498,16 @@ SSE/Admin 状态至少覆盖：
 
 本设计阶段完成条件：
 
-1. 本文件写入 `docs/superpowers/specs/`。
+1. 本文件写入 `docs/specs/`。
 2. 本文件通过占位、矛盾、范围和模糊性自检。
 3. `TASKS.md` 记录第三期设计上下文。
 4. API 和根仓库文档变更分别提交。
 5. 用户 review 本文件。
-6. 用户确认后进入 Superpowers `writing-plans`，再产出可执行计划。
+6. 用户确认后进入 KT workflow `KT plan writing`，再产出可执行计划。
 
-## 交接到 writing-plans
+## 交接到 KT plan writing
 
-`writing-plans` 必须把本设计转换为可执行计划，至少包含：
+`KT plan writing` 必须把本设计转换为可执行计划，至少包含：
 
 - Batch 0-8 逐批任务清单。
 - 每批 RED/GREEN 测试入口。

docs/specs/2026-06-15-qqbot-architecture-convergence-bangdream-rewrite-design.md

@@ -380,7 +380,7 @@ API caller 保持第三期规划：
 7. 拆分 NapCat Runtime 到第三期模块结构。
 8. Admin 同步拆 caller 和页面状态。
 9. 删除旧目录、旧 facade、旧 registry、旧大桶 shared、旧 Nest plugin services。
-10. 运行本地验证、全局 review、Superpowers review，再进入上线闭环。
+10. 运行本地验证、全局 review、KT global review，再进入上线闭环。
 
 ## 完成标准
 

docs/specs/2026-06-16-qqbot-plugin-scheduled-tasks-design.md

@@ -386,4 +386,4 @@ Admin TDD：
 
 实现计划应拆成后端平台能力、BangDream 任务、Admin 页面和上线闭环四段。每段都必须有独立 RED/GREEN 检查，不能先写 UI 再补接口，也不能让 BangDream 直接绕过平台调度。
 
-本 spec 通过后，下一步进入 Superpowers `writing-plans`。
+本 spec 通过后，下一步进入 KT workflow `KT plan writing`。

docs/specs/2026-06-18-admin-environment-dashboard-design.md

@@ -944,7 +944,7 @@ EnvironmentDashboardPage
 
 ### 设计验收
 
-- 用户确认本 spec 后，进入 Superpowers writing-plans。
+- 用户确认本 spec 后，进入 KT workflow KT plan writing。
 - 实施计划必须拆 API 合同、API 聚合、远程只读 adapter、Admin 页面、测试/线上闭环五部分。
 
 ### API 验收

docs/specs/2026-06-23-qqbot-napcat-source-fork-qr-refresh-design.md

@@ -334,7 +334,7 @@ docker rm -f kt-napcat-v3-verify
 ## 上线顺序
 
 1. 在 API 仓库提交本设计文档。
-2. 进入 `superpowers:writing-plans`，把源码 fork、KT 镜像集成、测试和线上 canary 拆成实施任务。
+2. 进入 `KT plan writing`，把源码 fork、KT 镜像集成、测试和线上 canary 拆成实施任务。
 3. 创建或更新 `D:\MyFiles\KT\GitHub\NapCatQQ` fork 分支。
 4. 先实现并验证 NapCatQQ 源码修复。
 5. 再修改 API 仓库的派生镜像构建脚本和测试。

docs/specs/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-design.md

@@ -147,7 +147,7 @@ The KT workspace on NAS must be a real full checkout, not a copy of temporary la
 
 Codex configuration parity is logical parity, not a byte-for-byte copy of the Windows profile:
 
-- Copy or generate non-secret configuration for trusted KT projects, model defaults, approval policy, sandbox policy, Superpowers skills, custom KT skills, and `ktWorkflow` MCP.
+- Copy or generate non-secret configuration for trusted KT projects, model defaults, approval policy, sandbox policy, KT-local skills, custom KT skills, and `ktWorkflow` MCP.
 - Convert Windows paths in `config.toml` to NAS Linux paths.
 - Keep GUI-only integrations such as local browser, computer-use, and Windows Figma local context disabled unless a NAS-compatible backend is explicitly installed.
 - Because the NAS is fully trusted, sensitive Codex files may be copied from the laptop when needed. The sync must still use a direct trusted channel, keep owner-only permissions, and exclude those files from Git, Jenkins artifacts, Docker build contexts, and public logs.

docs/specs/2026-06-24-napcatqq-upstream-sync-runtime-release-pipeline-design.zh-CN.md

@@ -147,7 +147,7 @@ NAS 上的 KT workspace 必须是真实全量 checkout，不是笔记本临时
 
 Codex 配置一致指“逻辑一致”，不是把 Windows profile 原样复制到 Linux：
 
-- 同步或生成非敏感配置：KT trusted projects、model 默认值、approval policy、sandbox policy、Superpowers skills、自定义 KT skills、`ktWorkflow` MCP。
+- 同步或生成非敏感配置：KT trusted projects、model 默认值、approval policy、sandbox policy、KT-local skills、自定义 KT skills、`ktWorkflow` MCP。
 - 把 `config.toml` 里的 Windows 路径转换成 NAS Linux 路径。
 - 本地浏览器、computer-use、Windows Figma local context 这类 GUI-only 能力默认禁用，除非后续安装了 NAS 可用 backend。
 - NAS 是完全可信环境，需要时可以从笔记本复制敏感 Codex 文件。同步仍必须走可信直连通道，文件 owner-only 权限，并排除在 Git、Jenkins artifact、Docker build context 和公开日志之外。