@xulthekl/team-flow 0.29.0 → 0.29.2

package/.claude/always/phase-guard.md

@@ -1,3 +1,3 @@
-# team-flow v0.29.0 | 阶段: {{state}} | 工作流: {{workflow}}
+# team-flow v0.29.2 | 阶段: {{state}} | 工作流: {{workflow}}
 当前阶段允许的操作由 workflow-start 路由规则定义。
 禁止跨越 DP gate 进入下一阶段。变更范围以 execution-contract.md 的 Intent Lock 为准。

package/.claude-plugin/marketplace.json

@@ -9,7 +9,7 @@
     {
       "name": "team-flow",
       "description": "8-state spec workflow + compound global compounding + architecture-design (4A/DDD) + local HTML prototype + product-level orchestration + bootstrap + e2e + session handoff + workflow feedback. 23 skills + 8 agents with embedded TDD, SDD, code review, debugging, delta spec sync, and design-system-driven prototyping.",
-      "version": "0.29.0",
+      "version": "0.29.2",
       "source": "./",
       "author": {
         "name": "LT",

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "team-flow",
-  "version": "0.29.0",
+  "version": "0.29.2",
   "description": "Unified workflow plugin: team-flow (spec-driven dev: TDD/SDD/code-review/debugging) + compound-engineering core subset (brainstorm/plan/compound/strategy/ideate/proof, global compounding) + architecture-design (4A+DDD incremental design & global compounding) + prototype (local HTML prototype, zero external deps) + e2e (Playwright E2E, AC-driven) + workflow-orchestrator (product-level workflow orchestration) + workflow-bootstrap (existing project onboarding) + session-handoff (context transfer) + workflow-feedback (issue tracking). 23 skills + 8 agents, one install.",
   "source": "./",
   "author": {

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "team-flow",
-  "version": "0.29.0",
+  "version": "0.29.2",
   "description": "Spec-first workflow that bridges OpenSpec-style planning and Superpowers-style execution discipline.",
   "author": {
     "name": "MageByte",

package/.cursor-plugin/marketplace.json

@@ -5,7 +5,7 @@
   },
   "metadata": {
     "description": "Unified workflow plugin marketplace for Cursor (team-flow: team-flow + compound + architecture-design + prototype).",
-    "version": "0.29.0"
+    "version": "0.29.2"
   },
   "plugins": [
     {

package/.cursor-plugin/plugin.json

@@ -2,7 +2,7 @@
   "name": "team-flow",
   "displayName": "team-flow",
   "description": "Unified workflow plugin: team-flow (spec-driven dev: TDD/SDD/code-review/debugging) + compound-engineering core subset (brainstorm/plan/compound/strategy/ideate/proof, global compounding) + architecture-design (4A+DDD incremental design & global compounding) + prototype (local HTML prototype, zero external deps) + e2e (Playwright E2E, AC-driven) + workflow-orchestrator (product-level workflow orchestration) + workflow-bootstrap (existing project onboarding) + session-handoff (context transfer) + workflow-feedback (issue tracking). 23 skills + 8 agents, one install.",
-  "version": "0.29.0",
+  "version": "0.29.2",
   "author": {
     "name": "LT",
     "url": "https://github.com/LT"

package/.github/plugin/marketplace.json

@@ -6,13 +6,13 @@
   },
   "metadata": {
     "description": "Unified workflow plugins and skills for AI coding agents (team-flow: team-flow + compound + architecture-design + prototype).",
-    "version": "0.29.0"
+    "version": "0.29.2"
   },
   "plugins": [
     {
       "name": "team-flow",
       "description": "Unified workflow with planning artifacts, execution contracts, TDD, review gates, systematic debugging, delta spec sync, architecture-design, and local HTML prototyping.",
-      "version": "0.29.0",
+      "version": "0.29.2",
       "source": ".",
       "author": {
         "name": "LT",

package/CHANGELOG.md

@@ -6,6 +6,14 @@ The format loosely follows Keep a Changelog.
 
 ## [Unreleased]
 
+## [0.29.1] - 2026-08-01
+
+### Fixed（子代理复用 SOP 强化）
+
+- **skills/workflow-start/SKILL.md**：四步协议摘要中明确"记录子代理 ID + 循环修正/DP-A 调整必须 SendMessage 恢复原子代理"
+- **skills/workflow-start/references/routing-rules.md**：Step 1 增加子代理 ID 记录指令；Step 2 审查循环从隐式 resume 改为"⛔ 必须 SendMessage 恢复，禁止启动新子代理"；Step 4d 从"优先 SendMessage"改为"⛔ 必须 SendMessage"；反模式清单增加"启动新子代理"禁令
+- 来源：LT 实测发现主代理在循环修正时启动新子代理而非恢复原子代理
+
 ## [0.29.0] - 2026-07-31
 
 ### Added（§37 DP-A 确认门 + 产物权限规则 + A4 增强，来源：workflow-feedback 2026-07-31 × 7）

package/GEMINI.md

@@ -8,7 +8,7 @@ The workflow is self-contained and does not require OpenSpec or Superpowers at r
 
 
 <!-- team-flow-phase-guard-start -->
-# team-flow v0.29.0 | 阶段: {{state}} | 工作流: {{workflow}}
+# team-flow v0.29.2 | 阶段: {{state}} | 工作流: {{workflow}}
 当前阶段允许的操作由 workflow-start 路由规则定义。
 禁止跨越 DP gate 进入下一阶段。变更范围以 execution-contract.md 的 Intent Lock 为准。
 <!-- team-flow-phase-guard-end -->

package/INSTALL.md

@@ -7,7 +7,7 @@
 - [Fission-AI/OpenSpec](https://github.com/Fission-AI/OpenSpec) — 规划引擎（Schema 验证、Delta Spec、工件解析）
 - [obra/superpowers](https://github.com/obra/superpowers) — 执行纪律（TDD 铁律、SDD、系统化调试、代码审查）
 
-当前发布版本：**v0.29.0**。
+当前发布版本：**v0.29.2**。
 
 ---
 

package/README.md

@@ -1,6 +1,6 @@
 # team-flow
 
-> 当前版本：`v0.29.0`
+> 当前版本：`v0.29.2`
 
 > 统一插件：**team-flow**（spec 驱动开发）+ **compound-engineering 核心子集**（全局复利）+ **architecture-design**（4A+DDD 增量设计）+ **prototype**（本地 HTML 原型）+ **e2e**（AC 驱动 E2E）+ **workflow-orchestrator**（产品级编排）+ **workflow-bootstrap**（既有项目接入）。一次安装，七套能力协同。
 

package/docs/README_en.md

@@ -126,7 +126,7 @@ npm install -g team-flow
 
 ### Version
 
-- Current: `v0.29.0`
+- Current: `v0.29.2`
 - v0.9.1 highlights: DP-4 execution-mode recommendations, a portable runtime across 17 platforms, and a raw-package smoke with no plugin-root variable.
 - Self-contained — no OpenSpec or Superpowers runtime required
 - Upstream: [Fission-AI/OpenSpec](https://github.com/Fission-AI/OpenSpec), [obra/superpowers](https://github.com/obra/superpowers)

package/gemini-extension.json

@@ -1,6 +1,6 @@
 {
   "name": "team-flow",
   "description": "Unified workflow plugin: team-flow (spec-driven dev) + compound-engineering core subset + architecture-design (4A/DDD) + prototype (local HTML). 17 skills, one install.",
-  "version": "0.29.0",
+  "version": "0.29.2",
   "contextFileName": "GEMINI.md"
 }

package/hooks/session-start

@@ -1,11 +1,11 @@
 #!/usr/bin/env bash
-# v0.29.0: auto-sync CLI version with plugin version
+# v0.29.2: auto-sync CLI version with plugin version
 set -e
 
 # ═══════════════════════════════════════════════════════════════
 # Plugin version (update this when releasing new versions)
 # ═══════════════════════════════════════════════════════════════
-PLUGIN_VERSION="0.29.0"
+PLUGIN_VERSION="0.29.2"
 
 # ═══════════════════════════════════════════════════════════════
 # Step 1: Auto-sync CLI version with plugin version

package/llms.txt

@@ -3,7 +3,7 @@
 ## Overview
 spec-superflow is a self-contained workflow integration plugin for Claude Code, Cursor, OpenAI Codex CLI/App, GitHub Copilot CLI, Gemini CLI, OpenCode, WorkBuddy, and Trae. It merges spec-driven planning artifacts (proposal, specs, design, tasks) with disciplined execution guardrails (TDD, review gates, controlled handoff) into one unified workflow.
 
-Current version: v0.29.0.
+Current version: v0.29.2.
 
 ## Key Documents
 - README.md: Chinese homepage with full usage guide and FAQ

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@xulthekl/team-flow",
-  "version": "0.29.0",
+  "version": "0.29.2",
   "description": "Unified plugin (23 skills + 10 agents) integrating team-flow, compound-engineering, architecture-design, prototype, design-system, workflow-orchestrator, workflow-bootstrap, e2e, session-handoff, workflow-feedback for multi-agent coding tools.",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",

package/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "team-flow",
-  "version": "0.29.0",
+  "version": "0.29.2",
   "description": "Unified workflow plugin: team-flow (spec-driven dev: TDD/SDD/code-review/debugging) + compound-engineering core subset (brainstorm/plan/compound/strategy/ideate/proof, global compounding) + architecture-design (4A+DDD incremental design & global compounding) + prototype (local HTML prototype, zero external deps) + e2e (Playwright E2E, AC-driven) + workflow-orchestrator (product-level workflow orchestration) + workflow-bootstrap (existing project onboarding) + session-handoff (context transfer) + workflow-feedback (issue tracking). 23 skills + 10 agents, one install.",
   "author": {
     "name": "LT"

package/skills/prototype/references/orchestration-flow.md

@@ -52,6 +52,12 @@ LT 确认后进入 ③。**confirmed_plan 序列化**：将 scout 简报的选
 
 派发插件 agent `prototype-builder`（有 Write），传入：已确认的原型方案（含选定方向）、`design-system.md` 路径、目标分支/版本、PRD 路径。
 
+**⛔ 子代理 ID 记录（v0.29.1，必须执行）**：dispatch 后**立即记录**子代理 ID，后续 ⑤ 循环修正和 ⑥ 人工评审调整**必须通过此 ID 恢复子代理**，不得启动新子代理：
+```
+builder_agent_id = <agentId>    # ③ dispatch 返回
+reviewer_agent_id = <agentId>   # ④ dispatch 返回
+```
+
 **prompt 组装顺序契约（v0.18.0, v0.24.0 扩展）**：`design-system.md → references/template.html → references/layouts.md → references/checklist.md → references/craft/（按页面类型选择注入）→ references/interactive-prototype.md（按需）→ references/wireframe.md（按需）→ PRD`。
 
 craft 注入策略：仪表盘→laws-of-ux+state-coverage；表单→state-coverage+accessibility-baseline；列表→typography-hierarchy；通用→anti-ai-slop（始终注入）。
@@ -91,7 +97,7 @@ test -f <prototype_path>/index.html && test -s <prototype_path>/index.html && ec
 
 ## 步骤 ⑤ 循环编排（主代理）
 
-- FAIL → 主代理把不一致项交给 prototype-builder 修正（回 ③），重新 ④ 评审，**≤3 轮**。
+- FAIL → **⛔ 必须通过 SendMessage 恢复原 prototype-builder 修正**（`SendMessage(to: builder_agent_id, message: "审查 FAIL，不一致项如下：{findings}。请修正后返回。")`），然后 **SendMessage 恢复原 prototype-reviewer 重新评审**（`SendMessage(to: reviewer_agent_id, message: "第 {N} 轮审查，原型已修正。请重新评审。")`），**≤3 轮**。**禁止启动新的 builder/reviewer 子代理**——原子代理拥有完整的设计上下文和审查历史。唯一例外：SendMessage 恢复失败时的 fallback。
 - **收敛检测**：连续两轮不一致项集合无缩小（修正无效/误报死循环）→ 立即转人工，不等满 3 轮。
 - PASS / PASS_WITH_WARNINGS → 进入 ⑥。
 
@@ -99,7 +105,7 @@ test -f <prototype_path>/index.html && test -s <prototype_path>/index.html && ec
 
 主代理用 AskUserQuestion 呈现评审结论 + 争议项，LT 选择：
 - **PRD 有问题** → 回 orchestrator S2 修订 PRD（**vN 内修订 + 变更履历，非升版**；见 feedback-loops 设计）→ PRD 更新后再回 ① 更新原型。**独立调用 fallback**（非 orchestrator 触发时）：无 S2 可回，直接提示 LT 修订 `prd/vN/prd.md` 后重入本 skill。
-- **原型需调整**（美观/体验/品牌/信息密度等自动评审查不到的维度）→ 回 ③ 派 prototype-builder 实施调整
+- **原型需调整**（美观/体验/品牌/信息密度等自动评审查不到的维度）→ **⛔ 必须通过 SendMessage 恢复原 prototype-builder 实施调整**（`SendMessage(to: builder_agent_id, message: "用户评审反馈：{调整意见}。请修改原型。")`），修改后 SendMessage 恢复原 prototype-reviewer 重新评审。**禁止启动新子代理**（唯一例外：SendMessage 恢复失败时的 fallback）
 - **通过** → 冻结：PRD frontmatter `frozen: true`（frozen_downstream），prototype 定版
 
 ## 子代理阻断处理（§18.1）

package/skills/workflow-start/SKILL.md

@@ -60,10 +60,10 @@ Guard: `arch_design_decision` in `.team-flow.yaml` is `null` → must run before
 
 **Four-step protocol (MUST execute in order)**:
 
-1. **Dispatch**: `architecture-design` as sub-agent → returns `decision` + `reason` + `artifacts`
-2. **Auto-review** (decision=required 时触发): 校验产物文件存在且非空 → dispatch `architecture-reviewer` sub-agent → FAIL 则循环修正（≤3 轮 + 收敛检测，不收敛转人工）→ 报告落盘 `changes/<name>/architecture/auto-review.md`
+1. **Dispatch**: `architecture-design` as sub-agent → returns `decision` + `reason` + `artifacts`。**⛔ 记录子代理 ID**（后续循环修正和 DP-A 调整必须通过此 ID 恢复，禁止启动新子代理）
+2. **Auto-review** (decision=required 时触发): 校验产物文件存在且非空 → dispatch `architecture-reviewer` sub-agent（**记录子代理 ID**）→ FAIL 则通过 **SendMessage 恢复原 architecture-design 子代理**修正（≤3 轮 + 收敛检测，不收敛转人工）→ 报告落盘 `changes/<name>/architecture/auto-review.md`
 3. **Reasonableness check + state write**: PASS/PASS_WITH_WARNINGS → write `arch_design_decision` + `arch_review_*` to yaml; skipped + brief 含架构关键词 → BLOCK; required + artifacts 缺失 → BLOCK; required + auto-review FAIL → BLOCK
-4. **DP-A 用户确认门（v0.29.0 §37）**: 输出架构决策摘要 → AskUserQuestion 确认 → 需要调整时转交子代理修改（禁止主代理直接修改）→ 修改后重新 auto-review → 回到本步骤重新确认。含项目规范变更提示（advisory）。详见 `references/routing-rules.md`「Step 4: DP-A」
+4. **DP-A 用户确认门（v0.29.0 §37）**: 输出架构决策摘要 → AskUserQuestion 确认 → 需要调整时**必须通过 SendMessage 恢复原子代理**修改（禁止主代理直接修改，禁止启动新子代理）→ 修改后 SendMessage 恢复原 reviewer 重新 auto-review → 回到本步骤重新确认。含项目规范变更提示（advisory）。详见 `references/routing-rules.md`「Step 4: DP-A」
 
 Full protocol in `references/routing-rules.md`「Route to architecture-design」.
 

package/skills/workflow-start/references/routing-rules.md

@@ -17,6 +17,12 @@ Dispatch `architecture-design` as sub-agent with inputs:
 - existing `specs/`
 - global `docs/architecture/`（As-Is 基线）
 
+**⛔ 子代理 ID 记录（v0.29.0 §37，必须执行）**：dispatch 后**立即记录**子代理 ID（Agent 工具返回的 `agentId` 或 task-notification 中的 `task-id`），后续 Step 2 审查循环和 Step 4 DP-A 调整**必须通过此 ID 恢复子代理**，不得启动新子代理。记录格式：
+```
+arch_design_agent_id = <agentId>   # Step 1 dispatch 返回
+arch_reviewer_agent_id = <agentId> # Step 2 dispatch 返回
+```
+
 Sub-agent returns structured output:
 ```yaml
 decision: required | skipped
@@ -52,6 +58,12 @@ reviewer 报告 → FAIL → 不一致项交给 architecture-design 修正 →
 reviewer 报告 → PASS / PASS_WITH_WARNINGS → 进入 Step 3
 ```
 
+**⛔ 循环修正必须通过 SendMessage 恢复原子代理（v0.29.0 §37，禁止启动新子代理）**：
+- FAIL 后修正：`SendMessage(to: arch_design_agent_id, message: "审查 FAIL，不一致项如下：{findings}。请修正后返回更新的结构化输出。")`
+- 重新审查：`SendMessage(to: arch_reviewer_agent_id, message: "第 {N} 轮审查，架构产物已修正。请重新执行 6 维度审查。")`
+- **禁止**在循环中启动新的 architecture-design 或 architecture-reviewer 子代理——原子代理拥有完整的设计上下文和审查历史，新子代理需要重新加载全部上下文（浪费 10-20K tokens 且可能丢失修正连贯性）
+- **唯一例外**：SendMessage 恢复失败（子代理 transcript 不可用）时，启动新子代理作为 fallback，但必须在 dispatch prompt 中传入：原产物路径 + auto-review 报告 + 历次修正记录
+
 报告落盘到 `changes/<name>/architecture/auto-review.md`。
 
 ### Step 3: Reasonableness check
@@ -148,12 +160,20 @@ AskUserQuestion:
 
 - **选择 B（需要调整）**：
   1. 记录用户调整意见
-  2. **通过子代理修改（禁止主代理直接 Edit 架构产物）**：
-     - 优先：`SendMessage(to: <原 architecture-design 子代理 ID>)` 发送修复指令，传入调整意见
-     - 子代理不可恢复时：启动新 architecture-design 子代理，传入：调整意见 + 原产物路径 + auto-review 报告
-  3. 子代理修改完成后，**重新触发 auto-review**（即使之前已 PASS）
+  2. **⛔ 必须通过 SendMessage 恢复原 architecture-design 子代理修改（禁止启动新子代理，禁止主代理直接 Edit）**：
+     ```
+     SendMessage(to: arch_design_agent_id, message: "用户评审反馈（DP-A 调整）：
+       {用户调整意见}
+       请修改架构产物，保持内部一致性，返回更新后的结构化输出。")
+     ```
+     原子代理拥有完整的设计上下文（PRD、brief、基线、conventions），能保持修改的一致性。启动新子代理会丢失上下文（重新加载 10-20K tokens）且可能引入不一致。
+  3. 子代理修改完成后，**通过 SendMessage 恢复原 architecture-reviewer 子代理重新审查**（即使之前已 PASS）：
+     ```
+     SendMessage(to: arch_reviewer_agent_id, message: "架构产物已按用户评审意见修正。请重新执行 6 维度审查。")
+     ```
   4. auto-review PASS 后，更新 yaml 状态（arch_review_verdict / arch_review_rounds）
   5. 回到 Step 4a 重新输出摘要 + 重新确认
+  6. **唯一 fallback**：SendMessage 恢复失败时，启动新子代理，但必须在 dispatch prompt 中传入：调整意见 + 原产物路径 + auto-review 报告 + 历次修正记录
   ```bash
   tf state set <change-dir> dp_a_result "adjustment_requested"
   tf state set <change-dir> dp_a_timestamp $(date -u +%Y-%m-%dT%H:%M:%SZ)
@@ -167,6 +187,7 @@ AskUserQuestion:
 - 主代理直接 Read + Edit `architecture/*.md` 文件（违反 Artifact Ownership 规则）
 - 跳过 DP-A 直接路由 spec-writer
 - 用户选择 B 后不重新 auto-review 就直接重新确认
+- **循环修正或 DP-A 调整时启动新的 architecture-design / architecture-reviewer 子代理**（必须通过 SendMessage 恢复原子代理，原子代理拥有完整上下文；新子代理 = 上下文断裂 + token 浪费。唯一例外：SendMessage 恢复失败时的 fallback）
 
 ## Route to spec-writer
 Guard: `tf runtime guard check <dir> exploring specifying --json` → fail = BLOCK.