xtep-workspace 1.10.10

package/plugin/skills/reviewing-spec/SKILL.md

@@ -0,0 +1,333 @@
+---
+name: reviewing-spec
+description: Use when performing Spec Review as dispatch_spec role to verify implementation completeness. Validates execution results against requirements and acceptance criteria.
+---
+
+# Reviewing Spec
+
+## Announcement (MANDATORY)
+
+Before executing this skill, you MUST announce to the user:
+
+「我将使用 **规格评审** 技能来验证实现完整性、检查是否满足需求和验收标准。」
+
+This creates a commitment checkpoint. Proceed only after announcing.
+
+## Core Thinking
+
+**Verify** - Systematically validate each acceptance criterion. Evidence-based verdicts only. Trust nothing - verify everything.
+
+## Role: dispatch_spec
+
+As a `dispatch_spec` node, you are part of the dispatch flow:
+- You are a **child** of the dispatch parent node
+- You verify the work done by the paired `dispatch_exec` node
+- You report back to the parent via `dispatch_complete`
+
+**Critical Principle**: You perform INDEPENDENT verification. You do NOT trust the exec node's conclusion. You verify against the original requirements and acceptance criteria by examining the actual code/output.
+
+## Typical Actions
+
+- Verify acceptance criteria (WHEN/THEN)
+- Check requirement coverage
+- Identify implementation gaps
+- Deliver pass/fail verdict
+- Report via dispatch_complete
+
+## SOP
+
+### 1. Gather Review Context
+
+**Goal**: Understand what needs to be reviewed.
+
+- **Target node**: Which execution node to review?
+- **Requirements**: What was the node supposed to do?
+- **Acceptance criteria**: What are the WHEN/THEN conditions?
+- **Execution conclusion**: What did the executor claim to complete?
+
+**Output**: Clear understanding of review scope
+
+### 2. Systematic Criterion Verification (Gate Function)
+
+**Goal**: INDEPENDENTLY verify each criterion by RUNNING verification commands.
+
+**Iron Law: NO PASS WITHOUT RUNNING VERIFICATION COMMANDS**
+
+For each criterion, check its **Verify** column and execute accordingly:
+
+#### If `[cmd]` - RUN the Command (MANDATORY)
+
+```bash
+# You MUST actually run the command, not assume it passes
+npm test -- --grep 'xxx'
+# or
+pytest tests/test_main.py
+# or whatever command is specified
+```
+
+1. **RUN** the command (FRESH - do not trust exec's output)
+2. **READ** the full output
+3. **CHECK** exit code (0 = success, non-zero = failure)
+4. **RECORD** evidence: `"[cmd] npm test → exit 0, 5/5 pass"`
+
+#### If `[manual]` - Verify Manual Steps Were Done
+
+1. **READ** exec's log for manual verification evidence
+2. **ASSESS** if steps were actually performed
+3. **RECORD** evidence: `"[manual] exec logged: 打开页面 → 组件显示正常"`
+
+#### If `[check]` - Perform Code Inspection
+
+1. **INSPECT** the specified target (file, function, pattern)
+2. **CONFIRM** condition is met
+3. **RECORD** evidence: `"[check] no TODO → grep found 0 matches"`
+
+#### If No Verify Column
+
+1. **INSPECT** code changes
+2. **ASSESS** if WHEN/THEN is logically satisfied
+3. **RECORD** reasoning (lower confidence)
+4. **⚠️ WARNING**: Flag as "unverified by command"
+
+**Evidence Table** (MUST produce):
+```markdown
+| # | Criterion | Verify | Command/Action | Result | Status |
+|---|-----------|--------|----------------|--------|--------|
+| 1 | WHEN... THEN... | [cmd] | npm test | exit 0, 3/3 | PASS |
+| 2 | WHEN... THEN... | [check] | grep TODO | 0 found | PASS |
+| 3 | WHEN... THEN... | [manual] | exec log | "UI checked" | PASS |
+```
+
+**Output**: Evidence table with command outputs for each criterion
+
+### 3. Requirement Coverage Check
+
+**Goal**: Verify all requirements are addressed.
+
+- **Completeness**: All aspects of requirement implemented?
+- **Correctness**: Implementation matches intent?
+- **Scope adherence**: No unauthorized additions/omissions?
+
+**Output**: Coverage assessment
+
+### 3.5. Implementation Completeness Check
+
+**Goal**: Verify no shortcuts or deferred work.
+
+**Search changed files for**:
+- `TODO`, `FIXME`, `HACK`, `XXX` comments
+- "暂时"、"临时"、"简化实现" in comments or code
+- Placeholder values or hardcoded data that should be configurable
+- Commented-out code with "later" notes
+
+**If found**: FAIL with specific locations and required fixes.
+
+**Output**: Completeness verification (PASS if no incomplete markers found)
+
+### 4. Compile Verdict
+
+**Goal**: Deliver final review result.
+
+- **PASS conditions**:
+  - ALL acceptance criteria passed
+  - Requirements fully covered
+  - No critical issues found
+
+- **FAIL conditions**:
+  - ANY acceptance criterion failed
+  - Missing required functionality
+  - Critical implementation issues
+  - TODO/FIXME/HACK markers found in code
+  - Simplified or temporary implementations detected
+
+**Output**: Verdict with detailed findings
+
+## Checklist
+
+### Context Gathering
+- [ ] Target execution node identified
+- [ ] Requirements understood
+- [ ] Acceptance criteria extracted
+- [ ] Execution conclusion reviewed
+
+### Criterion Verification (Gate Function)
+- [ ] Each `[cmd]` command ACTUALLY RUN (not trusted from exec)
+- [ ] Each `[manual]` step verified from exec's log
+- [ ] Each `[check]` inspection performed
+- [ ] Evidence Table produced with command outputs
+- [ ] Status recorded (pass/fail) with evidence
+
+### Coverage
+- [ ] All requirements addressed
+- [ ] No scope deviation
+- [ ] Implementation matches intent
+
+### Completeness
+- [ ] No TODO/FIXME/HACK markers in changed code
+- [ ] No "暂时/临时/简化" workarounds
+- [ ] No placeholder or hardcoded values
+- [ ] No deferred implementations
+
+### Verdict
+- [ ] Pass/fail determined
+- [ ] Findings documented
+- [ ] Actionable feedback provided (if fail)
+
+## Output Template
+
+```markdown
+### Spec Review Report
+
+**Target Node**: [node-id] - [title]
+**Review Node**: [review-node-id]
+**Verdict**: PASS / FAIL
+
+### Verification Evidence Table (MANDATORY)
+
+| # | Criterion | Verify | Command/Action | Result | Status |
+|---|-----------|--------|----------------|--------|--------|
+| 1 | WHEN [condition] THEN [result] | [cmd] | `npm test` | exit 0, 3/3 pass | PASS |
+| 2 | WHEN [condition] THEN [result] | [check] | `grep TODO` | 0 found | PASS |
+| 3 | WHEN [condition] THEN [result] | [manual] | exec log | "verified" | PASS |
+
+### Requirement Coverage
+
+- **Covered**: [list of covered aspects]
+- **Missing**: [list of missing aspects, if any]
+- **Scope Issues**: [deviations, if any]
+
+### Findings
+
+#### Issues (if any)
+1. [Issue description with evidence]
+2. [Issue description with evidence]
+
+#### Suggestions (if fail)
+1. [Specific action to fix issue 1]
+2. [Specific action to fix issue 2]
+
+### Conclusion
+
+[Summary of review result]
+```
+
+## Recording to Workspace
+
+**Principle**: Review results must be actionable.
+
+### What to Record
+
+| Content | Where | Example |
+|---------|-------|---------|
+| Verdict summary | conclusion | "Spec Review PASS: 3/3 criteria met" |
+| Failed criteria | conclusion | "FAIL: Criterion 2 not met - [reason]" |
+| Full report | conclusion or memo | Detailed findings |
+
+### Conclusion Template (PASS)
+
+```
+**Verdict**: PASS
+**Criteria**: [X]/[X] passed
+**Coverage**: Complete
+**Summary**: Implementation fully meets requirements.
+```
+
+### Conclusion Template (FAIL)
+
+```
+**Verdict**: FAIL
+**Criteria**: [X]/[Y] passed, [Z] failed
+**Failed**:
+- Criterion N: [reason]
+**Required Actions**:
+1. [specific fix needed]
+```
+
+---
+
+## Red Flags
+
+1. **Skip criteria** - Not checking all WHEN/THEN conditions
+2. **Trust exec's output** - Using exec's command output instead of running yourself
+3. **No evidence** - Passing without verifiable proof
+4. **Skip [cmd]** - Not actually running the verification command
+5. **Subjective judgment** - Opinion-based instead of evidence-based
+6. **No Evidence Table** - Verdict without structured evidence
+7. **Partial pass** - Passing when some criteria fail
+8. **Ignore incomplete markers** - Passing code with TODO/FIXME/HACK
+9. **Accept shortcuts** - Passing simplified implementations
+
+## Mandatory Rules
+
+1. **MUST verify INDEPENDENTLY** - NEVER trust exec's conclusion, verify yourself
+2. **MUST RUN verification commands** - `[cmd]` means YOU run it, not trust exec ran it
+3. **MUST check ALL criteria** - Skipping any criterion is review failure
+4. **MUST provide evidence** - Every pass/fail needs verifiable proof (command output)
+5. **MUST produce Evidence Table** - No table = incomplete review
+6. **MUST fail if ANY criterion fails** - Partial pass is not pass
+7. **NEVER be lenient** - Pass only when ALL criteria are met
+8. **MUST check for incomplete markers** - TODO/FIXME/HACK in code = automatic FAIL
+9. **MUST reject simplified implementations** - "暂时/临时/简化" workarounds = FAIL
+
+## Anti-Patterns
+
+| Pattern | Wrong | Right |
+|---------|-------|-------|
+| **Assumption** | "Looks complete" | Check each criterion explicitly |
+| **Leniency** | Pass despite missing feature | Fail with clear feedback |
+| **Vague feedback** | "Needs improvement" | "Criterion 2 fails: no error handling in line 45" |
+| **Scope creep** | Review code style in spec review | Focus only on requirements/criteria |
+| **Trust exec blindly** | "Exec said it's done, so pass" | Verify independently against criteria |
+
+## Common Rationalizations
+
+| Excuse | Why Wrong | Correct Action |
+|--------|-----------|----------------|
+| "Exec's conclusion looks thorough" | Exec may have missed something | Verify independently |
+| "Exec already ran the test" | You need FRESH evidence, exec may have lied/erred | RUN the command yourself |
+| "I can see exec's output in the log" | Logs can be fabricated or outdated | RUN and see output yourself |
+| "It mostly works, close enough" | Partial implementation = partial pass = FAIL | ALL criteria must pass |
+| "The missing part is minor" | Criteria exist for a reason | Fail and specify what's missing |
+| "Code looks good, must work" | Looking good != working correctly | RUN verification command |
+| "I'll be lenient this time" | Leniency erodes quality standards | Standards apply equally every time |
+
+---
+
+## Completing the Review
+
+**IMPORTANT**: As a dispatch_spec node, you MUST call `dispatch_complete` to finalize your review.
+
+### On Pass
+
+```typescript
+dispatch_complete({
+  workspaceId: "...",
+  nodeId: "[your-node-id]",
+  success: true,
+  conclusion: `**Verdict**: PASS
+**Criteria**: [X]/[X] passed
+**Coverage**: Complete
+**Summary**: Implementation fully meets requirements.`
+})
+```
+
+### On Fail
+
+```typescript
+dispatch_complete({
+  workspaceId: "...",
+  nodeId: "[your-node-id]",
+  success: false,
+  conclusion: `**Verdict**: FAIL
+**Criteria**: [X]/[Y] passed, [Z] failed
+**Failed**:
+- Criterion N: [specific reason with evidence]
+**Required Actions**:
+1. [specific fix needed]`
+})
+```
+
+**Why dispatch_complete?**
+- Signals completion to the dispatch parent
+- Enables Git commit (on success) or rollback (on failure)
+- Provides structured feedback for retry decisions

package/plugin/skills/starting-info-flow/SKILL.md

@@ -0,0 +1,196 @@
+---
+name: starting-info-flow
+description: Use when user requests research, analysis, or summary in an existing workspace. Guides capability-based info node creation.
+---
+
+# Starting Info Flow
+
+## Announcement (MANDATORY)
+
+Before executing this skill, you MUST announce to the user:
+
+「我将使用 **信息流启动** 技能来引导能力选择、创建信息收集/总结节点。」
+
+This creates a commitment checkpoint. Proceed only after announcing.
+
+## When to Use
+
+User says things like:
+- "我想实现一下 xxx"
+- "帮我调研一下 xxx"
+- "分析一下这个问题"
+- "总结一下 xxx"
+- "研究一下 xxx 的实现"
+- "整理一下目前的讨论"
+
+In an **existing workspace**, when user wants to start a new info collection or summary task.
+
+## Core Flow
+
+```
+User requests research/analysis
+    ↓
+capability_list()  ← Get recommended capabilities
+    ↓
+Show user and ask for selection
+    ↓
+capability_select(infoType=?, selected=[...])  ← Create info node + capability children
+    ↓
+Execute capability tasks (call corresponding Skill)
+```
+
+## SOP
+
+### 1. Identify Task Type
+
+| User Request | infoType |
+|--------------|----------|
+| 调研、探索、研究、分析问题、排查 | `info_collection` |
+| 总结、整理、提炼、归纳、梳理 | `info_summary` |
+
+**Decision rule**:
+- Need to **actively explore** codebase/docs → `info_collection`
+- Need to **organize existing info** from conversation → `info_summary`
+
+### 2. Call capability_list
+
+Get capabilities for the workspace scenario:
+
+```typescript
+capability_list({ workspaceId: "ws-xxx" })  // Uses workspace's scenario
+```
+
+**Returns**:
+- `basePack`: Required capabilities (**ALL must be selected**)
+- `optionalPack`: Optional capabilities (user chooses)
+
+### 3. Show User and Ask for Selection
+
+**Template**:
+```
+要进行「{用户目标}」，这是一个**{收集/总结}**任务。
+
+我将执行以下能力：
+- **{能力名}**: {在此任务语境下的具体作用}
+- **{能力名}**: {在此任务语境下的具体作用}
+
+还可以选择：
+  a. **{能力名}** - {语境下描述}
+  b. **{能力名}** - {语境下描述}
+
+需要添加哪些？（输入序号如 a、ab，或回车跳过）
+```
+
+**Rules**:
+- basePack: 直接说"我将执行"，不问用户
+- optionalPack: 用 **abc 序号**，方便用户输入
+- 描述: **根据任务语境定制**，不用通用描述
+- **MUST** wait for user response before capability_select
+
+### 4. Call capability_select
+
+```typescript
+capability_select({
+  workspaceId: "ws-xxx",
+  infoType: "info_collection",  // or "info_summary"
+  selected: ["intent_alignment", "context_discovery", ...]
+})
+```
+
+**Returns**: `actionRequired` with Skill list for each capability.
+
+### 5. Execute Capability Tasks (CRITICAL)
+
+**capability_select 返回 `actionRequired`，MUST 遵循其指示调用 Skill。**
+
+```
+capability_select 返回
+    ↓
+actionRequired: { type: "invoke_skill", skills: [...] }
+    ↓
+Skill(aligning-intent)  ← 第一个能力
+    ↓
+node_transition(action="start")
+    ↓
+按 Skill SOP 执行
+    ↓
+node_transition(action="complete")
+    ↓
+Skill(discovering-context)  ← 下一个能力
+    ↓
+...重复...
+```
+
+**执行规则**：
+
+| 步骤 | 动作 | 说明 |
+|------|------|------|
+| 1 | `Skill(xxx)` | 调用对应 Skill 获取 SOP |
+| 2 | `node_transition(action="start")` | 启动节点 |
+| 3 | 按 SOP 执行 | 遵循 Skill 的 SOP 步骤 |
+| 4 | `node_transition(action="complete")` | 完成节点 |
+
+**⚠️ 禁止事项**：
+- **NEVER** 跳过 Skill 调用直接 node_transition start
+- **NEVER** 在信息收集/总结阶段修改任何代码
+- **NEVER** 在能力节点中执行 Write/Edit/Update 操作
+
+**信息收集/总结阶段的职责边界**：
+
+| 允许 | 禁止 |
+|------|------|
+| Read、Search、Grep、Glob | Write、Edit、Update |
+| 分析代码、理解结构 | 修改代码、创建文件 |
+| 记录发现到 conclusion | 直接实现功能 |
+| 提出方案建议 | 执行方案 |
+
+**为什么必须调用 Skill？**
+- Skill 包含该能力的 SOP（标准操作流程）
+- Skill 定义了正确的执行步骤和检查清单
+- 不读 Skill = 不知道该能力的正确做法
+
+## Mandatory Rules
+
+1. **MUST call capability_list first** - Get available capabilities for this scenario
+2. **MUST show user and ask** - Never decide capabilities alone
+3. **MUST use capability_select** - Never node_create info nodes directly
+4. **MUST include all basePack** - selected must contain all basePack capabilities
+5. **MUST choose correct infoType** - Based on task nature (collect vs summarize)
+6. **MUST call Skill before node_transition start** - 每个能力节点启动前必须先调用对应 Skill
+7. **NEVER modify code in info phase** - 信息收集/总结阶段禁止 Write/Edit/Update
+8. **MUST follow Skill SOP** - Skill 定义了正确的执行流程，必须遵循
+
+## Red Flags
+
+1. **Direct node_create for info tasks** - Should use capability flow
+2. **Skip user confirmation** - Decide capabilities without asking
+3. **Wrong infoType** - Using collection when summary is appropriate
+4. **Partial basePack** - Missing required capabilities from basePack
+5. **Skip capability Skill** - Complete capability_select but not call the Skill
+6. **Skip Skill, direct node_transition** - 没调用 Skill 就直接启动节点 ⚠️ 严重错误
+7. **Write/Edit in info phase** - 在信息收集/总结阶段修改代码 ⚠️ 严重错误
+8. **Ignore actionRequired** - capability_select 返回的 actionRequired 被忽略
+
+## Anti-Patterns
+
+| Wrong | Right |
+|-------|-------|
+| `node_create(role="info_collection")` | `capability_list` → `capability_select` |
+| Decide all capabilities yourself | Show user and ask for selection |
+| Always use info_collection | Match infoType to task nature |
+| Skip the returned Skill list | Call each Skill in order |
+| node_transition start without Skill | Skill(xxx) first, then node_transition |
+| Write/Edit in info_collection node | Only Read/Search, record to conclusion |
+| Ignore actionRequired from capability_select | Follow actionRequired instructions exactly |
+
+## Common Rationalizations
+
+| Excuse | Why Wrong | Correct Action |
+|--------|-----------|----------------|
+| "User already knows what they want" | User may not know all capability options | Show capabilities and ask |
+| "node_create is faster" | Loses Skill guidance and structure | Use capability flow |
+| "info_collection covers everything" | Wrong type wastes exploration effort | Match task nature |
+| "I'll just start working" | Skipping capability flow loses guidance | Follow the SOP |
+| "I know what to do, don't need Skill" | Skill 定义了正确流程，你可能遗漏关键步骤 | 必须先调用 Skill 获取 SOP |
+| "Just a quick fix while exploring" | 信息收集阶段禁止修改，会破坏职责边界 | 记录发现，创建执行节点后再修改 |
+| "The change is obvious, no need for execution node" | 所有修改必须在执行节点中进行，确保可追踪 | 创建执行节点，按流程执行 |

package/web/README.md

@@ -0,0 +1,5 @@
+# Vue 3 + TypeScript + Vite
+
+This template should help get you started developing with Vue 3 and TypeScript in Vite. The template uses Vue 3 `<script setup>` SFCs, check out the [script setup docs](https://v3.vuejs.org/api/sfc-script-setup.html#sfc-script-setup) to learn more.
+
+Learn more about the recommended Project Setup and IDE Support in the [Vue Docs TypeScript Guide](https://vuejs.org/guide/typescript/overview.html#project-setup).