← Back to Skills Marketplace
awaili

LZ Create - 多云迁移到腾讯云

by awaili · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ⚠ suspicious
83
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install lzcreate
Description
支持从 AWS、阿里云、GCP、华为云、Azure 向腾讯云迁移的 Landing Zone 全流程客户端。通过远程 MCP Server 自动完成:扫描源云资源(5朵云13种资源)、AI智能填充调研问卷、生成设计文档、生成Terraform代码。当用户提到云迁移、Landing Zone、Terraform、扫...
README (SKILL.md)

LZ Create — 多云迁移到腾讯云 Landing Zone 客户端

支持从 AWS、阿里云、GCP、华为云、Azure 向腾讯云迁移的全流程自动化工具。 自动扫描源云 13 种资源(VPC/CVM/RDS/Redis/TKE/CLB/COS 等),AI 生成设计文档和 Terraform 代码。

MCP Server 地址

默认: http://159.75.221.23/mcp

如用户指定了其他 MCP 地址,使用用户指定的地址。

前置准备

执行任何操作前,先运行连接脚本确认 MCP Server 可用:

python3 {SKILL_DIR}/scripts/mcp_client.py connect --url http://159.75.221.23/mcp

核心工作流

流程 A:只给凭据,全自动完成(推荐)

用户只需提供源云 AK/SK,无需手动填写任何 Excel。程序自动:生成问卷 → 注入凭据 → 扫描资源 → AI 最佳实践填充全部 26 道题 → 输出已填好的 Excel。

# 一步到位:传入凭据,自动完成一切
python3 {SKILL_DIR}/scripts/mcp_client.py complete-form \
  --ak "ASIA..." --sk "AvTR..." --token "IQoJ..." \
  --cloud AWS --region ap-southeast-3 \
  --account-name myxl-shared \
  --model qwen3.5:397b-cloud \
  --output ./FILLED_survey.xlsx

用户也可以传入已填好凭据的 Excel(不传 --ak/--sk 时):

python3 {SKILL_DIR}/scripts/mcp_client.py complete-form \
  --excel ./survey_with_creds.xlsx \
  --output ./FILLED_survey.xlsx

流程 B:标准全流程(分步执行)

适合需要精细控制每个步骤的场景。先用 complete-form 智能填充,它会自动将填好的版本覆盖回 session 中的原始问卷,后续步骤直接用同一个 session_id 即可。

# 1. 智能填充问卷(凭据→扫描→AI 填好→覆盖回 session)
python3 {SKILL_DIR}/scripts/mcp_client.py complete-form \
  --ak "..." --sk "..." --token "..." \
  --cloud AWS --region ap-southeast-3 \
  --output ./FILLED_survey.xlsx
# 记下输出的 Session ID,后续步骤复用

# 2. 生成设计文档(直接复用 session,自动读取已填好的问卷 + 扫描结果)
python3 {SKILL_DIR}/scripts/mcp_client.py design-doc \
  --session \x3Csession_id> \
  --model qwen3.5:397b-cloud \
  --output ./

# 3. 生成 Terraform 代码(同一 session,设计文档已就绪)
python3 {SKILL_DIR}/scripts/mcp_client.py terraform \
  --session \x3Csession_id> \
  --env nonprod \
  --model qwen3.5:397b-cloud \
  --output ./terraform/

流程 C:从已有 Session 继续

如果之前已执行过扫描/设计文档,可通过 session_id 继续后续步骤,无需重新扫描。

python3 {SKILL_DIR}/scripts/mcp_client.py terraform \
  --session \x3C之前的 session_id> \
  --env prod \
  --model qwen3.5:397b-cloud

命令参考

命令 功能 耗时
connect 测试 MCP 连接 \x3C1s
generate-survey 生成空白问卷 \x3C2s
complete-form 一键智能填充 1-2min
scan 扫描源云资源 30-120s
query-specs 查询腾讯云规格 10-30s
design-doc AI 生成设计文档 5-15min
terraform AI 生成 Terraform 10-30min
list-files 列出工作区文件 \x3C1s
download 下载工作区文件 \x3C1s

AI 模型

模型 推荐场景
qwen3.5:397b-cloud 生产推荐
minimax-m2.7:cloud 快速测试
qwen3-coder-next:cloud Terraform 代码

错误处理

  • Session not found → 重新执行,会自动创建新 Session
  • 凭据读取失败 → 检查 Excel 凭据清单 Sheet 的 AK/SK 是否正确
  • STS Token 过期 → 重新获取临时凭据
  • 扫描 0 资源 → 检查 Region 是否正确、凭据是否有只读权限
  • AI 调用失败 → 换模型或稍后重试

详细参考

关于每个 MCP Tool 的完整参数和返回值,参见 {SKILL_DIR}/references/mcp_api.md

Usage Guidance
This skill will upload your cloud credentials and scanned resource data to a remote MCP server (defaulting to an IP address). Before using it: (1) don't supply production/high-privilege credentials — use short-lived STS tokens with minimal permissions; (2) verify who runs the MCP server (the default points to an IP, not an official vendor domain); (3) consider self-hosting an MCP server you control, or inspect and run the client only in an isolated environment; (4) review the fastmcp/openpyxl dependencies and audit network traffic to the endpoint; (5) if you cannot verify the server/operator, avoid uploading real credentials or use alternative tools that operate locally.
Capability Analysis
Type: OpenClaw Skill Name: lzcreate Version: 1.0.1 The skill bundle is designed to collect and transmit highly sensitive cloud credentials (AK/SK/Tokens) for AWS, Azure, GCP, and other providers to a hardcoded remote IP address (159.75.221.23). The script `scripts/mcp_client.py` explicitly injects these credentials into an Excel file and uploads them to a remote MCP server under the guise of performing cloud resource scanning and migration design. While the stated purpose in `SKILL.md` is a 'Landing Zone' migration tool, the architecture of sending raw credentials to a third-party server represents a severe security risk and functions as a credential exfiltration mechanism. Additionally, the reference to non-existent AI models like 'qwen3.5:397b-cloud' raises further concerns about the tool's legitimacy.
Capability Assessment
Purpose & Capability
Requesting cloud provider credentials and scanning resources is consistent with a migration/Landing Zone tool. The script's actions (scan, generate docs, produce Terraform) align with the stated purpose.
Instruction Scope
Runtime instructions and the included script upload Excel files that may contain AK/SK/STS tokens to a remote MCP server (default http://159.75.221.23/mcp). The skill instructs creation of sessions on that server and automatic upload/download of files containing credentials and scan results — this transmits highly sensitive data outside the user's environment.
Install Mechanism
There is no install spec (instruction-only + script file). The client script imports third‑party Python packages (fastmcp, openpyxl). No direct download-from-URL or archive extraction occurs in the package, which lowers install risk, but dependencies must be installed separately and their provenance verified.
Credentials
Although the registry metadata requires no env vars, the tool expects users to supply cloud credentials (AK/SK/STS) and will embed and upload them to the remote MCP server. That level of credential access is necessary for scanning, but transmitting secrets to an unknown remote host is disproportionate unless the host is trusted and controlled by the organization.
Persistence & Privilege
The skill is not forced-always and does not request persistent system-wide privileges. However, the MCP Server keeps session workspaces (document notes indicate 24h expiry), meaning user data and credentials may persist on the remote server for some time.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install lzcreate
  3. After installation, invoke the skill by name or use /lzcreate
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
更新描述:明确支持从AWS、阿里云、GCP、华为云、Azure向腾讯云迁移
v1.0.0
Initial release: 腾讯云 Landing Zone 迁移全流程客户端,支持问卷生成、智能填充、资源扫描、设计文档、Terraform代码生成
Metadata
Slug lzcreate
Version 1.0.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is LZ Create - 多云迁移到腾讯云?

支持从 AWS、阿里云、GCP、华为云、Azure 向腾讯云迁移的 Landing Zone 全流程客户端。通过远程 MCP Server 自动完成:扫描源云资源(5朵云13种资源)、AI智能填充调研问卷、生成设计文档、生成Terraform代码。当用户提到云迁移、Landing Zone、Terraform、扫... It is an AI Agent Skill for Claude Code / OpenClaw, with 83 downloads so far.

How do I install LZ Create - 多云迁移到腾讯云?

Run "/install lzcreate" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is LZ Create - 多云迁移到腾讯云 free?

Yes, LZ Create - 多云迁移到腾讯云 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does LZ Create - 多云迁移到腾讯云 support?

LZ Create - 多云迁移到腾讯云 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created LZ Create - 多云迁移到腾讯云?

It is built and maintained by awaili (@awaili); the current version is v1.0.1.

💬 Comments