Autoops Standard Skill

AutoOps 标准技能（完整版）

本技能是该项目的"标准化总技能入口"，提供从代码到部署再到验证的完整 SOP 流程。

目标与承诺

• 完整性：原始规范文档保存在 references 中
• 一致性：提供校验脚本验证规范未漂移
• 可执行：完整 SOP（audit → build → runtime → test → integration）
• 可验证：包含 curl 测试、MCP Playwright 浏览器测试、README 功能覆盖检查

参考材料

按以下顺序加载：

统一 SOP（总览）

┌─────────────────────────────────────────────────────────────────┐
│                    AutoOps 标准化 SOP                            │
├─────────────────────────────────────────────────────────────────┤
│ 1. 校验规范一致性 (validate_skill.sh)                            │
│ 2. 项目审计 (Audit Gate)                                         │
│ 3. 构建容器 (Build Gate)                                         │
│ 4. 启动服务 (Runtime Gate)                                       │
│ 5. curl API 测试 (API Gate)                                      │
│ 6. MCP Playwright 浏览器测试 (UI Gate)                           │
│ 7. README 功能覆盖验证 (Feature Gate)                            │
│ 8. 安全与可观测性检查 (Security/Observability Gate)              │
│ 9. 集成到 Homepage (Integration Gate)                            │
└─────────────────────────────────────────────────────────────────┘

质量门禁

阶段 1：校验规范一致性

bash skills/autoops-standard-skill/scripts/validate_skill.sh

阶段 2：项目审计

输出 audit_report.yaml：

audit_result:
  status: PASS|WARN|FAIL
  tool_slug: "<tool_slug>"
  detected:
    language: "Python|Node|Go"
    mode: "cli|web|library"
  requirements:
    network: true
    browser: false
    gpu: false
  risks: []

阶段 3：构建 Docker 容器

# 构建镜像
./scripts/deploy-<tool>-web.sh build

# 验证镜像存在
docker images | grep "<tool>-web"

阶段 4：curl API 测试（构建后必做）

4.1 启动容器

./scripts/deploy-<tool>-web.sh run
# 等待 5 秒让服务启动
sleep 5

4.2 健康检查测试

# 基础健康检查
curl -fsS http://localhost:<PORT>/health
# 期望输出: {"status": "healthy"} 或 {"status": "ok"}

# 带超时的健康检查
curl -fsS --max-time 10 http://localhost:<PORT>/health || echo "FAIL: health check failed"

4.3 API 端点测试

# 1. 工具信息 API
curl -fsS http://localhost:<PORT>/api/info | jq .
# 验证: 返回 JSON 包含 name, version, description

# 2. 任务启动 API
TASK_RESPONSE=$(curl -fsS -X POST http://localhost:<PORT>/api/process \
  -H 'Content-Type: application/json' \
  -d '{"input": "test", "options": {}}')
TASK_ID=$(echo "$TASK_RESPONSE" | jq -r '.task_id')
echo "Task ID: $TASK_ID"

# 3. 任务状态查询
curl -fsS http://localhost:<PORT>/api/tasks/${TASK_ID} | jq .

# 4. 获取语音/选项列表 (如适用)
curl -fsS http://localhost:<PORT>/api/voices | jq . 2>/dev/null || true

4.4 curl 测试脚本模板

#!/bin/bash
# test-api.sh - API 自动化测试脚本

set -e

TOOL_NAME="<tool>"
PORT="${PORT:-5400}"
BASE_URL="http://localhost:${PORT}"

log_pass() { echo -e "\033[32m[PASS]\033[0m $1"; }
log_fail() { echo -e "\033[31m[FAIL]\033[0m $1"; exit 1; }

echo "=== Testing ${TOOL_NAME} API ==="

# Test 1: Health check
if curl -fsS --max-time 5 "${BASE_URL}/health" > /dev/null 2>&1; then
    log_pass "Health check"
else
    log_fail "Health check"
fi

# Test 2: API info
INFO=$(curl -fsS "${BASE_URL}/api/info" 2>/dev/null)
if echo "$INFO" | jq -e '.name' > /dev/null 2>&1; then
    log_pass "API info - name: $(echo $INFO | jq -r '.name')"
else
    log_fail "API info"
fi

# Test 3: Process endpoint exists
if curl -fsS -X POST "${BASE_URL}/api/process" \
     -H 'Content-Type: application/json' \
     -d '{}' 2>&1 | grep -q "task_id\|error"; then
    log_pass "Process endpoint responds"
else
    log_fail "Process endpoint"
fi

echo "=== All API tests passed ==="

阶段 5：MCP Playwright 浏览器测试

使用 MCP Playwright 进行自动化浏览器测试，验证 Web UI 功能。

5.1 Playwright 测试场景

5.2 MCP Playwright 测试流程

使用 MCP Playwright 执行以下测试:

1. 导航到 http://localhost:<PORT>
2. 等待页面加载完成 (检查关键元素存在)
3. 截图保存为 test-screenshots/01-homepage.png
4. 定位表单输入框，输入测试内容
5. 点击提交按钮
6. 等待任务完成或结果显示
7. 截图保存为 test-screenshots/02-result.png
8. 验证结果区域内容
9. 如有下载按钮，验证下载链接有效

5.3 Playwright 测试检查清单

• 页面在 5 秒内加载完成
• 无 console.error 错误
• 主要交互元素可见且可点击
• 表单提交后有响应反馈
• 任务状态实时更新 (WebSocket)
• 结果可下载或可查看
• 移动端视口下布局正常

5.4 Playwright 测试报告模板

playwright_test_report:
  tool: "<tool_slug>"
  url: "http://localhost:<PORT>"
  timestamp: "2026-01-28T10:00:00Z"
  tests:
    - name: "Page Load"
      status: PASS|FAIL
      duration_ms: 1234
      screenshot: "test-screenshots/01-homepage.png"
    - name: "Form Submission"
      status: PASS|FAIL
      duration_ms: 2345
      screenshot: "test-screenshots/02-submit.png"
    - name: "Result Display"
      status: PASS|FAIL
      duration_ms: 3456
      screenshot: "test-screenshots/03-result.png"
  console_errors: []
  overall: PASS|FAIL

阶段 6：README 功能覆盖验证

验证 UI 实现的功能是否与 README/tool.yaml 文档一致。

6.1 功能对照矩阵

对于每个工具，创建功能对照表：

6.2 README 功能提取流程

# 从 docs/<tool>-web-ui-guide.md 提取功能列表
grep -E "^- \*\*|^### " docs/<tool>-web-ui-guide.md

# 从 tool.yaml 提取功能描述
yq '.long_description' tools/<tool>-main/tool.yaml
yq '.api_endpoints[].description' tools/<tool>-main/tool.yaml

6.3 功能验证清单生成

#!/bin/bash
# generate-feature-checklist.sh

TOOL_NAME="$1"
TOOL_YAML="tools/${TOOL_NAME}-main/tool.yaml"
GUIDE_MD="docs/${TOOL_NAME}-web-ui-guide.md"

echo "=== ${TOOL_NAME} 功能验证清单 ==="
echo ""

echo "## tool.yaml 声明的功能:"
if [ -f "$TOOL_YAML" ]; then
    yq '.long_description' "$TOOL_YAML" 2>/dev/null | grep -E "^  - " || echo "(无)"
fi

echo ""
echo "## API 端点:"
if [ -f "$TOOL_YAML" ]; then
    yq '.api_endpoints[].endpoint + " - " + .api_endpoints[].description' "$TOOL_YAML" 2>/dev/null || echo "(无)"
fi

echo ""
echo "## 使用指南中的功能:"
if [ -f "$GUIDE_MD" ]; then
    grep -E "^- \*\*" "$GUIDE_MD" || echo "(无)"
fi

6.4 功能遗漏报告模板

feature_coverage_report:
  tool: "<tool_slug>"
  source_docs:
    - "docs/<tool>-web-ui-guide.md"
    - "tools/<tool>-main/tool.yaml"
  total_features: 10
  implemented: 9
  missing: 1
  coverage: "90%"
  details:
    implemented:
      - "文本转语音"
      - "多语言支持"
      - "语速调节"
    missing:
      - feature: "批量处理"
        documented_in: "tool.yaml long_description"
        priority: "low"
        notes: "可在后续版本添加"

阶段 7：安全与可观测性检查

7.1 安全检查

# 检查硬编码密钥
grep -rn "password\|secret\|api_key\|token" tools/<tool>-main/<tool>-web/ --include="*.py" | grep -v "environ\|getenv"

# 检查是否支持环境变量配置
grep -n "os.environ\|os.getenv" tools/<tool>-main/<tool>-web/app.py

7.2 日志检查

# 检查容器日志无 fatal 错误
docker logs <tool>-web-container 2>&1 | tail -n 200 | \
  grep -E "Traceback|panic:|Segmentation fault|FATAL|CRITICAL" && \
  echo "FAIL: Fatal errors found" || echo "PASS: No fatal errors"

# 检查服务启动日志
docker logs <tool>-web-container 2>&1 | \
  grep -E "Starting|Listening|Server started|port" | head -5

阶段 8：集成验证

8.1 Homepage 集成检查

# 确认 tool.yaml 存在且格式正确
cat tools/<tool>-main/tool.yaml | yq '.container_name'

# 确认容器名符合规范
docker ps --format "{{.Names}}" | grep "<tool>-web-container"

# 通过 Homepage API 检查工具是否被发现
curl -fsS http://localhost:8000/api/tools | jq '.[] | select(.name | contains("<tool>"))'

8.2 集成测试清单

• tool.yaml 的 container_name 与实际容器名一致
• 容器名以 -container 结尾
• Homepage 可列出该工具
• 可通过 Homepage 启动/停止该工具
• 工具卡片显示正确的状态

完整测试脚本

#!/bin/bash
# full-test.sh - 完整测试流程

set -e

TOOL_NAME="${1:-edge-tts}"
PORT="${2:-5500}"

echo "========================================="
echo "AutoOps 工具完整测试: ${TOOL_NAME}"
echo "========================================="

# 1. 构建
echo "[1/6] Building..."
./scripts/deploy-${TOOL_NAME}-web.sh build

# 2. 启动
echo "[2/6] Starting container..."
./scripts/deploy-${TOOL_NAME}-web.sh run
sleep 5

# 3. curl API 测试
echo "[3/6] API tests..."
curl -fsS http://localhost:${PORT}/health || exit 1
curl -fsS http://localhost:${PORT}/api/info | jq . || exit 1

# 4. 日志检查
echo "[4/6] Log check..."
docker logs ${TOOL_NAME}-web-container 2>&1 | tail -50 | \
  grep -E "Traceback|FATAL" && exit 1 || true

# 5. MCP Playwright 测试 (需要 MCP 环境)
echo "[5/6] Browser test..."
echo "  -> 请使用 MCP Playwright 执行浏览器测试"
echo "  -> 目标 URL: http://localhost:${PORT}"

# 6. 功能覆盖检查
echo "[6/6] Feature coverage..."
echo "  -> 请对照 docs/${TOOL_NAME}-web-ui-guide.md 检查功能"

echo "========================================="
echo "测试完成！请手动验证 Playwright 和功能覆盖"
echo "========================================="

test_report.json 完整模板

{
  "tool_slug": "<tool_slug>",
  "timestamp": "2026-01-28T10:00:00Z",
  "gates": {
    "audit": {"status": "PASS", "notes": ""},
    "build": {"status": "PASS", "image": "<tool>-web:latest"},
    "runtime": {"status": "PASS", "container": "<tool>-web-container", "port": 5500},
    "api": {
      "status": "PASS",
      "tests": [
        {"endpoint": "/health", "method": "GET", "status": 200},
        {"endpoint": "/api/info", "method": "GET", "status": 200},
        {"endpoint": "/api/process", "method": "POST", "status": 200}
      ]
    },
    "ui": {
      "status": "PASS",
      "playwright_tests": [
        {"name": "page_load", "status": "PASS"},
        {"name": "form_submit", "status": "PASS"},
        {"name": "result_display", "status": "PASS"}
      ],
      "screenshots": ["01-homepage.png", "02-result.png"]
    },
    "feature": {
      "status": "PASS",
      "coverage": "100%",
      "total": 10,
      "implemented": 10,
      "missing": []
    },
    "security": {"status": "PASS", "hardcoded_secrets": false},
    "observability": {"status": "PASS", "fatal_errors": []},
    "integration": {"status": "PASS", "homepage_visible": true}
  },
  "overall": "PASS"
}

快速命令参考

# 校验规范
bash skills/autoops-standard-skill/scripts/validate_skill.sh

# 构建并测试
./scripts/deploy-<tool>-web.sh build
./scripts/deploy-<tool>-web.sh run
curl http://localhost:<PORT>/health

# 查看日志
docker logs <tool>-web-container

# 停止服务
./scripts/deploy-<tool>-web.sh stop

注意事项

• 本技能不覆盖下载型插件/安装包的 SOP，请参见：
  • skills/plugin-distribution-standardizer/SKILL.md
• 本技能聚焦 deployable/service standardization
• MCP Playwright 测试需要配置 Playwright MCP 环境
• 功能覆盖验证依赖 docs/<tool>-web-ui-guide.md 文档

Converted and distributed by TomeVault — claim your Tome and manage your conversions.