create-harness-vibe-coding 0.1.3 → 0.1.4

package/templates/common/docs/harness/lifecycle.md

@@ -0,0 +1,33 @@
+# 0-1 Product Lifecycle
+
+Use when starting a new product, clarifying a vague idea, or deciding the next phase.
+
+## Phase Contract
+
+| Phase | Input | Output | Gate |
+| --- | --- | --- | --- |
+| Idea | user intent | problem, target user, non-goals | unclear points asked or assumptions recorded |
+| Research | problem and constraints | `research/research-results.md` | `research/README.md` followed; at least 3 references or explicit reason not possible |
+| PRD | research decision | `research/PRD.md` | MVP, non-goals, acceptance criteria are verifiable |
+| Architecture | PRD | `harness/architecture.md`, `domain/ports.md` | boundaries and first ports are defined |
+| Plan | PRD and architecture | `docs/harness/PLAN.md`, optional `docs/harness/dispatch.md`, one `docs/features/<name>.md` per PRD scope item | tasks have owners, write sets, verification |
+| Build | plan and tests | minimal vertical slice | tests or manual checks prove behavior |
+| Verify | implementation | review findings, test evidence | no unresolved critical/high findings |
+| Feedback | verified slice | next iteration or release decision | learnings recorded in PRD, PLAN, or MEMORY |
+
+## Operating Rules
+
+- Move one phase at a time unless the user explicitly asks for a fast lane.
+- Start coding only after PRD and minimum architecture gates pass.
+- Prefer one thin vertical slice over broad scaffolding.
+- If feedback changes scope, update PRD before implementation.
+- If implementation reveals a boundary problem, update architecture or ports before continuing.
+
+## Fast Lane
+
+Small edits may skip full lifecycle when all are true:
+
+- user intent is clear
+- one file or one narrow behavior
+- no architecture, port, data-flow, state, permission, or public API change
+- one verification command or one manual check is enough

package/templates/common/docs/harness/state-machines.md

@@ -1,23 +1,23 @@
-# 状态机 — {{projectName}}
+# State Machines — {{projectName}}
 
-> **职责**：定义所有有状态组件的状态、转移、守卫条件、边界行为。长流程 harness 的大量 bug 来自非法状态转移。
+> **Responsibility**: Define the states, transitions, guard conditions, and boundary behaviors for every stateful component. The majority of bugs in long-running harness pipelines originate from illegal state transitions.
 >
-> 哲学来源：UML 2.5.1 第 15.3.14 节 + "转移表是整个文档最重要的产物"。
+> Philosophical source: UML 2.5.1 Section 15.3.14 + "The transition table is the single most important artifact in the document."
 
 ---
 
-## 状态机模板
+## State Machine Template
 
-每个有状态组件按此格式填写：
+Fill out each stateful component using this format:
 
-### 状态枚举
+### State Enumeration
 
-| 状态 | 描述 | 进入条件 | 退出条件 |
+| State | Description | Entry Condition | Exit Condition |
 | --- | --- | --- | --- |
 | `{{STATE_1}}` | {{DESCRIPTION}} | {{CONDITION}} | {{CONDITION}} |
 | `{{STATE_2}}` | {{DESCRIPTION}} | {{CONDITION}} | {{CONDITION}} |
 
-### 状态转移图
+### State Transition Diagram
 
 ```mermaid
 stateDiagram-v2
@@ -26,25 +26,25 @@ stateDiagram-v2
     {{STATE_2}} --> {{INITIAL_STATE}} : {{TRIGGER}}
 ```
 
-### 转移表（最重要）
+### Transition Table (Most Important)
 
-| 当前状态 ↓ / 事件 → | `{{EVENT_1}}` | `{{EVENT_2}}` | `{{EVENT_3}}` |
+| Current State ↓ / Event → | `{{EVENT_1}}` | `{{EVENT_2}}` | `{{EVENT_3}}` |
 | --- | --- | --- | --- |
 | **`{{STATE_1}}`** | {{TARGET}} | {{TARGET}} | {{TARGET}} |
 | **`{{STATE_2}}`** | {{TARGET}} | {{TARGET}} | {{TARGET}} |
 
-### 守卫条件
+### Guard Conditions
 
-| 转移 | 守卫条件 | 备注 |
+| Transition | Guard Condition | Notes |
 | --- | --- | --- |
 | `{{SOURCE}} -> {{TARGET}}` | {{GUARD}} | {{NOTE}} |
 
-### 非法转移
+### Illegal Transitions
 
-| 转移 | 为什么非法 |
+| Transition | Why Illegal |
 | --- | --- |
 | `{{SOURCE}} -> {{TARGET}}` | {{REASON}} |
 
 ---
 
-> **提示**：当前 state-machines.md 是空模板。请按项目实际的有状态组件填写。
+> **Note**: The current state-machines.md is an empty template. Please fill it in according to the project's actual stateful components.

package/templates/common/docs/research/PRD.md

@@ -1,63 +1,65 @@
 # PRD: {{projectName}}
 
-> **读者**：AI + 未来的你。不是给 PM 审批用的，是给实现和审查用的。
-> **原则**：不超过一页。用 checkbox，不用散文。负面定义 > 正面定义。
+> **Audience**: AI + future you. Not for PM approval — for implementation and review.
+> **Principle**: One page max. Use checkboxes, not prose. Negative definitions > positive definitions.
 >
-> 哲学来源：Miqdad Jaffer (OpenAI) 的 AI 时代精益 PRD 模板。
+> Philosophy source: Miqdad Jaffer (OpenAI)'s lean PRD template for the AI era.
 
 ---
 
-## 1. 为什么做
+## 1. Why
 
 {{WHY_THIS_PROJECT_EXISTS}}
 
-## 2. MVP 边界
+## 2. MVP Scope
 
-### v0.1 必须能做
+### v0.1 Must Be Able To
+
+> Each item below must be split into a separate `docs/features/<name>.md` created from `docs/features/_template.md` before implementation begins. One feature = one doc = one implementation unit.
 
 - [ ] {{MUST_1}}
 - [ ] {{MUST_2}}
 - [ ] {{MUST_3}}
 
-### 明确不做
+### Explicitly Out of Scope
 
 - {{NON_GOAL_1}}
 - {{NON_GOAL_2}}
 - {{NON_GOAL_3}}
 
-## 3. 决策优先级
+## 3. Decision Priorities
 
-1. **安全边界** — 不允许绕过权限、安全门和审计。
-2. **架构边界** — 不允许让 domain 依赖 harness/infrastructure/interfaces，也不让 harness 承担领域判断。
-3. **可验证性** — 新行为必须有测试或明确的人工验证记录。
-4. **正确性** — 行为符合设计文档、端口合同、数据流和状态机。
-5. **简洁性** — 最少代码解决当前 MVP，不为未来场景预埋抽象。
+1. **Security Boundary** — No bypassing permissions, security gates, or audit.
+2. **Architecture Boundary** — Domain must not depend on harness/infrastructure/interfaces, and harness must not make domain judgments.
+3. **Verifiability** — New behavior must have tests or explicit manual verification records.
+4. **Correctness** — Behavior must match design docs, port contracts, data flows, and state machines.
+5. **Simplicity** — Minimum code to solve the current MVP. No abstractions pre-built for future scenarios.
 
-## 4. 用户与使用场景
+## 4. Users & Usage Scenarios
 
-| 用户角色 | 核心场景 | 频率 | 痛点 |
+| User Role | Core Scenario | Frequency | Pain Point |
 | --- | --- | --- | --- |
 | {{USER_ROLE_1}} | {{SCENARIO}} | {{FREQUENCY}} | {{PAIN}} |
 | {{USER_ROLE_2}} | {{SCENARIO}} | {{FREQUENCY}} | {{PAIN}} |
 
-## 5. 验收标准
+## 5. Acceptance Criteria
 
 - [ ] {{ACCEPTANCE_1}}
 - [ ] {{ACCEPTANCE_2}}
 - [ ] {{ACCEPTANCE_3}}
 
-## 6. 非功能性需求
+## 6. Non-Functional Requirements
 
-| 维度 | 目标 | 测量方式 |
+| Dimension | Target | Measurement |
 | --- | --- | --- |
 | {{DIMENSION_1}} | {{TARGET}} | {{MEASUREMENT}} |
 | {{DIMENSION_2}} | {{TARGET}} | {{MEASUREMENT}} |
 
 ---
 
-## 填充完成标准
+## Fill Completion Standard
 
-- [ ] MVP 和 Non-goals 都是项目事实，不保留 `{{...}}` 占位符。
-- [ ] 决策优先级能指导 tradeoff，而不是泛泛而谈。
-- [ ] 每条验收标准都能被测试、命令或人工步骤验证。
-- [ ] 如果实现偏离 PRD，先更新本文件再改代码。
+- [ ] MVP and Non-goals are project facts; no `{{...}}` placeholders remain.
+- [ ] Decision priorities guide tradeoffs, not generic platitudes.
+- [ ] Every acceptance criterion is verifiable by test, command, or manual step.
+- [ ] If implementation diverges from the PRD, update this file before changing code.

package/templates/common/docs/research/README.md

@@ -0,0 +1,121 @@
+# Research Protocol
+
+Purpose: guide agent research. Record conclusions in [research-results.md](research-results.md); do not put raw search notes there.
+
+## When To Research
+
+Use this before PRD, stack choice, external API use, public dependency choice, pricing/legal/security assumptions, or any fact likely to change.
+
+Skip only when the task is local, narrow, and fully answerable from existing project files. Record the skip reason in `docs/harness/PLAN.md`.
+
+## Research Agent
+
+Spawn a Research Agent, or emulate one in a bounded pass, when any are true:
+
+- more than two external sources are needed
+- the answer depends on current docs, releases, pricing, policy, or community practice
+- a framework, API, architecture, or dependency choice affects implementation
+
+Built-in agents:
+
+- `.claude/agents/researcher.md`: product, market, open-source, dependency, pricing, policy, and ecosystem research.
+- `.claude/agents/docs-researcher.md`: official docs, API, SDK, config, limits, errors, and examples verification.
+
+For multi-agent research plus build work, create the dispatch table in `docs/harness/PLAN.md` and follow `docs/harness/dispatch.md`.
+
+Research Agent input:
+
+- question
+- decision needed
+- source boundaries
+- allowed tools
+- output format
+
+Research Agent output:
+
+- searched queries and tools used
+- source list with links and source type
+- adopted / rejected / watch decisions
+- risks, unknowns, and follow-up questions
+- patch-ready updates for [research-results.md](research-results.md)
+
+## Tool Order
+
+1. Local first: inspect this repo, existing docs, lockfiles, tests, and package metadata.
+2. GitHub / open source: prefer official repos, docs folders, examples, issues with maintainer answers, releases, and active forks.
+3. Tavily: use when configured for broad web search or source discovery.
+4. TinyFish: use when configured for rendered pages, dynamic sites, browser workflows, or structured extraction.
+5. Built-in web search: free fallback when no external research tool is configured; expect less structure and verify more carefully.
+6. Ask the user for sources when network or tool access is unavailable.
+
+Do not require paid tools. If Tavily or TinyFish is unavailable, use the fallback and record the limitation.
+
+## Optional Tool Setup
+
+Tavily:
+
+```text
+# Optional: requires a Tavily API key or configured CLI/tool.
+tvly search "query" --depth advanced --max-results 10 --json
+tvly search "query" --include-domains github.com,docs.github.com --json
+```
+
+TinyFish:
+
+```bash
+# Optional: requires TINYFISH_API_KEY or tinyfish auth login.
+npm install -g @tiny-fish/cli
+tinyfish auth login
+tinyfish agent run --url "https://example.com" "Extract product data. Return JSON."
+```
+
+GitHub CLI:
+
+```bash
+gh search repos "topic keywords" --archived=false --json fullName,url,description,stargazersCount,pushedAt
+gh search code "symbol or config" --repo owner/name
+```
+
+Fallback web search examples:
+
+```text
+site:github.com <framework> starter template
+site:github.com <library> examples
+site:docs.<vendor>.com <api> limits errors
+<product category> alternatives pricing docs
+```
+
+## Source Rules
+
+- Prefer primary sources: official docs, official GitHub repos, standards, papers, release notes.
+- Use community sources for pitfalls and adoption signals only; label them as community evidence.
+- Check dates for unstable facts.
+- Compare at least three sources, or record why fewer are enough.
+- Do not copy large source text. Summarize the decision-relevant facts.
+
+## Requirement Quality
+
+Use these patterns when turning research into PRD or feature docs:
+
+- PRD: why, target user, MVP, non-goals, success measures, acceptance criteria.
+- EARS: `When <trigger>, the <system> shall <response>` for precise requirements.
+- Gherkin: `Given / When / Then` for testable behavior scenarios.
+- Spec-first: requirements before plan, plan before tasks, tasks before implementation.
+
+## Write Target
+
+- Research process, queries, and limitations: this file or `docs/harness/PLAN.md`.
+- Final research decisions: [research-results.md](research-results.md).
+- Product scope: [PRD.md](PRD.md).
+- Architecture consequences: `docs/harness/architecture.md` and `docs/domain/ports.md`.
+
+## Method References
+
+- GitHub Spec Kit: spec-first phases and AI coding-agent workflow: https://github.github.com/spec-kit/
+- Atlassian PRD guidance: goals, assumptions, user stories, out-of-scope, success criteria: https://www.atlassian.com/agile/requirements
+- EARS: structured textual requirements: https://alistairmavin.com/ears/
+- Cucumber Gherkin: `Given / When / Then` executable examples: https://cucumber.io/docs/gherkin/reference
+- Tavily search docs: search depth, domain filters, max results, raw content: https://docs.tavily.com/documentation/api-reference/endpoint/search
+- Tavily CLI docs: `tvly search`, crawl, map, extract: https://docs.tavily.com/documentation/tavily-cli
+- TinyFish CLI docs: web search and browser-agent runs: https://docs.tinyfish.ai/cli
+- GitHub CLI search docs: repository and code search from terminal: https://cli.github.com/manual/gh_search_repos

package/templates/common/docs/research/research-results.md

@@ -0,0 +1,66 @@
+# {{projectName}} - Research Results
+
+> **Purpose**: Record final research results and decisions. Use [README.md](README.md) for process, tools, queries, and research-agent rules.
+> **Principle**: Each candidate has a clear Purpose / Strength / Weakness / Decision.
+
+---
+
+## Research Date
+
+{{YYYY-MM-DD}}
+
+## Research Goal
+
+{{WHAT_YOU_ARE_RESEARCHING}}
+
+---
+
+## Candidate References
+
+### 1) {{CANDIDATE_1}}
+
+- **What it is**: {{DESCRIPTION}}
+- **Source Type**: {{OFFICIAL_DOCS / GITHUB_REPO / COMMUNITY / PAPER / OTHER}}
+- **Checked Date**: {{YYYY-MM-DD}}
+- **Strengths**: {{STRENGTHS}}
+- **Weaknesses**: {{WEAKNESSES}}
+- **Decision**: {{ADOPT / REJECT / WATCH}}
+- **Link**: {{URL}}
+
+### 2) {{CANDIDATE_2}}
+
+- **What it is**: {{DESCRIPTION}}
+- **Source Type**: {{OFFICIAL_DOCS / GITHUB_REPO / COMMUNITY / PAPER / OTHER}}
+- **Checked Date**: {{YYYY-MM-DD}}
+- **Strengths**: {{STRENGTHS}}
+- **Weaknesses**: {{WEAKNESSES}}
+- **Decision**: {{ADOPT / REJECT / WATCH}}
+- **Link**: {{URL}}
+
+### 3) {{CANDIDATE_3}}
+
+- **What it is**: {{DESCRIPTION}}
+- **Source Type**: {{OFFICIAL_DOCS / GITHUB_REPO / COMMUNITY / PAPER / OTHER}}
+- **Checked Date**: {{YYYY-MM-DD}}
+- **Strengths**: {{STRENGTHS}}
+- **Weaknesses**: {{WEAKNESSES}}
+- **Decision**: {{ADOPT / REJECT / WATCH}}
+- **Link**: {{URL}}
+
+---
+
+## Final Decision
+
+- **Architecture Style**: {{DECISION}}
+- **Core References**: {{REFERENCES}}
+- **Key Constraints**:
+  - {{CONSTRAINT_1}}
+  - {{CONSTRAINT_2}}
+  - {{CONSTRAINT_3}}
+
+---
+
+## Not Adopted But Worth Watching
+
+- {{ALTERNATIVE_1}}
+- {{ALTERNATIVE_2}}

package/templates/common/scripts/validate-harness.mjs

@@ -0,0 +1,207 @@
+#!/usr/bin/env node
+import fs from 'node:fs';
+import path from 'node:path';
+
+const root = process.cwd();
+const args = new Set(process.argv.slice(2));
+const strict = args.has('--strict') || args.has('--post-bootstrap');
+
+if (args.has('--help') || args.has('-h')) {
+  console.log(`Usage: node scripts/validate-harness.mjs [--strict]
+
+Default mode checks scaffold structure, links, agents, and skills.
+--strict also fails when project fact docs still contain {{...}} placeholders.`);
+  process.exit(0);
+}
+
+const commonAgents = [
+  'researcher',
+  'docs-researcher',
+  'planner',
+  'architect',
+  'test-writer',
+  'implementer',
+  'debugger',
+  'reviewer',
+  'verifier',
+];
+
+const commonSkills = [
+  'harness-router',
+  'harness-lifecycle',
+  'harness-research',
+  'harness-context',
+  'harness-build-loop',
+];
+
+const required = [
+  'AGENTS.md',
+  'CLAUDE.md',
+  'MEMORY.md',
+  '.claude/settings.json',
+  '.claude/rules/ecc/common.md',
+  ...commonAgents.map(agent => `.claude/agents/${agent}.md`),
+  ...commonSkills.map(skill => `.claude/skills/${skill}/SKILL.md`),
+  'docs/README.md',
+  'docs/harness/PLAN.md',
+  'docs/harness/lifecycle.md',
+  'docs/harness/dispatch.md',
+  'docs/harness/extension.md',
+  'docs/harness/context-loading.md',
+  'docs/harness/agent-workflow.md',
+  'docs/harness/architecture.md',
+  'docs/harness/data-flow.md',
+  'docs/harness/state-machines.md',
+  'docs/features/_template.md',
+  'docs/research/README.md',
+  'docs/research/research-results.md',
+  'docs/research/PRD.md',
+  'docs/domain/ports.md',
+];
+
+const projectFacts = [
+  'docs/harness/PLAN.md',
+  'docs/research/PRD.md',
+  'docs/research/research-results.md',
+  'docs/harness/architecture.md',
+  'docs/domain/ports.md',
+];
+
+const contextPacks = [
+  'Explorer Pass:',
+  'Planner:',
+  'Researcher:',
+  'Docs Researcher:',
+  'Architect:',
+  'Test Writer:',
+  'Implementer:',
+  'Reviewer:',
+  'Debugger:',
+  'Verifier:',
+];
+
+const errors = [];
+
+function read(rel) {
+  const file = path.join(root, rel);
+  return fs.existsSync(file) ? fs.readFileSync(file, 'utf8') : '';
+}
+
+function requireText(rel, text, label = text) {
+  const body = read(rel);
+  if (body && !body.includes(text)) errors.push(`${rel} missing ${label}`);
+}
+
+function frontmatterField(text, field) {
+  const match = text.match(new RegExp(`^${field}:\\s*(.+)$`, 'm'));
+  return match ? match[1].trim() : '';
+}
+
+for (const rel of required) {
+  if (!fs.existsSync(path.join(root, rel))) {
+    errors.push(`missing required file: ${rel}`);
+  }
+}
+
+if (fs.existsSync(path.join(root, 'docs/research/scaffolds.md'))) {
+  errors.push('legacy research file should be renamed: docs/research/scaffolds.md -> docs/research/research-results.md');
+}
+
+if (strict) {
+  for (const rel of projectFacts) {
+    const text = read(rel);
+    if (text.includes('{{')) {
+      errors.push(`template placeholders remain in project fact file: ${rel}`);
+    }
+  }
+}
+
+const docsReadme = read('docs/README.md');
+if (docsReadme) {
+  for (const marker of ['## Keyword Routing', '## Load By Task', 'When to Read', 'Keywords']) {
+    if (!docsReadme.includes(marker)) errors.push(`docs/README.md missing router marker: ${marker}`);
+  }
+}
+
+const plan = read('docs/harness/PLAN.md');
+if (plan) {
+  for (const heading of ['## Current Goal', '## Phase', '## Success Criteria', '## Loaded Context', '## Tasks', '## Parallel Dispatch', '## Verification']) {
+    if (!plan.includes(heading)) errors.push(`docs/harness/PLAN.md missing heading: ${heading}`);
+  }
+}
+
+const dispatch = read('docs/harness/dispatch.md');
+if (dispatch) {
+  for (const agent of commonAgents) {
+    if (!dispatch.includes(`\`${agent}\``)) errors.push(`docs/harness/dispatch.md missing common agent: ${agent}`);
+  }
+  if (!dispatch.includes('## Handoff Format')) errors.push('docs/harness/dispatch.md missing heading: ## Handoff Format');
+}
+
+const contextLoading = read('docs/harness/context-loading.md');
+if (contextLoading) {
+  if (!contextLoading.includes('docs/README.md` is the primary router')) {
+    errors.push('docs/harness/context-loading.md must declare docs/README.md as the primary router');
+  }
+  for (const pack of contextPacks) {
+    if (!contextLoading.includes(pack)) errors.push(`docs/harness/context-loading.md missing subagent pack: ${pack}`);
+  }
+}
+
+const memory = read('MEMORY.md');
+if (memory) {
+  for (const agent of commonAgents) {
+    const rel = `.claude/agents/${agent}.md`;
+    if (!memory.includes(rel)) errors.push(`MEMORY.md missing agent registration: ${rel}`);
+  }
+  for (const skill of commonSkills) {
+    const rel = `.claude/skills/${skill}/SKILL.md`;
+    if (!memory.includes(rel)) errors.push(`MEMORY.md missing skill registration: ${rel}`);
+  }
+}
+
+for (const skill of commonSkills) {
+  const rel = `.claude/skills/${skill}/SKILL.md`;
+  const text = read(rel);
+  if (!text) continue;
+  if (frontmatterField(text, 'name') !== skill) errors.push(`${rel} frontmatter name does not match directory`);
+  if (!frontmatterField(text, 'description')) errors.push(`${rel} missing frontmatter field: description`);
+}
+
+for (const agent of commonAgents) {
+  const rel = `.claude/agents/${agent}.md`;
+  const text = read(rel);
+  if (!text) continue;
+
+  for (const field of ['name', 'description', 'tools', 'model', 'skills']) {
+    if (!frontmatterField(text, field)) errors.push(`${rel} missing frontmatter field: ${field}:`);
+  }
+
+  if (frontmatterField(text, 'name') !== agent) {
+    errors.push(`${rel} frontmatter name does not match filename`);
+  }
+
+  const skill = frontmatterField(text, 'skills');
+  if (skill && !commonSkills.includes(skill)) {
+    errors.push(`${rel} references unknown skill: ${skill}`);
+  }
+
+  if (skill && !fs.existsSync(path.join(root, `.claude/skills/${skill}/SKILL.md`))) {
+    errors.push(`${rel} references missing skill file: .claude/skills/${skill}/SKILL.md`);
+  }
+}
+
+requireText('docs/harness/extension.md', 'Skills should extend the harness');
+requireText('docs/harness/agent-workflow.md', 'docs/harness/PLAN.md');
+requireText('docs/research/README.md', 'research-results.md');
+
+if (errors.length) {
+  console.error(`Harness validation failed${strict ? ' (strict)' : ''}:`);
+  for (const error of errors) console.error(`- ${error}`);
+  process.exit(1);
+}
+
+console.log(`Harness validation passed${strict ? ' (strict)' : ''}.`);
+if (!strict) {
+  console.log('Tip: run `node scripts/validate-harness.mjs --strict` after bootstrap to check unresolved project placeholders.');
+}

package/templates/common/docs/research/scaffolds.md

@@ -1,60 +0,0 @@
-# {{projectName}} — 前期调研
-
-> **职责**：记录"为什么选了这些技术/参考"。新加入的人不需要重新做一遍调研。
-> **原则**：每个候选物有明确的 Purpose / Strength / Weakness / Decision。
-
----
-
-## 调研时间
-
-{{YYYY-MM-DD}}
-
-## 调研目标
-
-{{WHAT_YOU_ARE_RESEARCHING}}
-
----
-
-## 候选参考物
-
-### 1) {{CANDIDATE_1}}
-
-- **是什么**：{{DESCRIPTION}}
-- **优势**：{{STRENGTHS}}
-- **局限**：{{WEAKNESSES}}
-- **结论**：{{ADOPT / REJECT / WATCH}}
-- **链接**：{{URL}}
-
-### 2) {{CANDIDATE_2}}
-
-- **是什么**：{{DESCRIPTION}}
-- **优势**：{{STRENGTHS}}
-- **局限**：{{WEAKNESSES}}
-- **结论**：{{ADOPT / REJECT / WATCH}}
-- **链接**：{{URL}}
-
-### 3) {{CANDIDATE_3}}
-
-- **是什么**：{{DESCRIPTION}}
-- **优势**：{{STRENGTHS}}
-- **局限**：{{WEAKNESSES}}
-- **结论**：{{ADOPT / REJECT / WATCH}}
-- **链接**：{{URL}}
-
----
-
-## 最终决策
-
-- **架构风格**：{{DECISION}}
-- **核心参考**：{{REFERENCES}}
-- **关键约束**：
-  - {{CONSTRAINT_1}}
-  - {{CONSTRAINT_2}}
-  - {{CONSTRAINT_3}}
-
----
-
-## 未采用但值得关注的
-
-- {{ALTERNATIVE_1}}
-- {{ALTERNATIVE_2}}