Norma by Quality Clouds

ai.qualityclouds由 qualityclouds.ai 發布官方發布者v1.0.1更新於 Oct 24, 2026

Checks AI-written code against your rules as it is written, and keeps a record of what it found.

已驗證Streamable HTTP可網頁執行開發者工具安全與監控

概覽

AI 產生的概覽

Norma by Quality Clouds 在 AI 撰寫程式的過程中依組織規則檢查程式碼,並留下檢查紀錄。

可以做什麼
  • 偵測儲存庫的技術堆疊並列出適用的規則集
  • 讀取某個規則集中的規則,包含 ID、嚴重程度、影響領域與修正指引
  • 依組織規則對單一檔案執行確定性的檢查
  • 列出已連結儲存庫上次完整掃描的未解決問題
  • 將已驗證合規、已修正與已預防的違規記錄到合規稽核軌跡
適用情境
當編碼代理在已連結至 Norma 帳戶的儲存庫中撰寫或審查程式碼,且希望每個檔案都依團隊標準檢查時使用;也適合處理上次完整掃描留下的問題並留下修正紀錄。
執行需求
遠端 Streamable HTTP 端點 api.qualityclouds.ai;首次連線需透過瀏覽器完成 OAuth 登入;需要 Norma 帳戶並透過 GitHub 或 Bitbucket 整合連結儲存庫,即時檢查僅能在已連結的儲存庫上執行;需要支援 MCP 的用戶端,例如 Cursor、Claude Code、Lovable 或 Replit。
安裝前請注意
OAuth 授權會讓伺服器存取你的 Norma 工作區,即時檢查只對已連結的儲存庫有效。每次檢查都會把檔案內容傳送到遠端服務,README 表示僅在檢查期間暫存於記憶體、不會儲存。register_applied_actions 會寫入合規稽核軌跡,應复核所記錄的結果。連結儲存庫即告知 Norma 該檢查屬於哪個儲存庫。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Norma by Quality Clouds,將其新增到工作區。
  2. 連接 所需的 憑證。它們會被私密儲存,絕不會在公開頁面上展示。
  3. 為需要使用其工具的對話啟用該服務。

Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式,或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南,先向你展示發布者、端點和登入方式,經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/ai.qualityclouds%2Fnorma/install.md,並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前,先告訴我它的發布者、端點和登入方式,等我確認。不要把 API 金鑰寫進命令或我會提交的檔案;問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。 客戶端會在首次連接時完成 OAuth 登入。

{
  "mcpServers": {
    "norma-by-quality-clouds": {
      "type": "http",
      "url": "https://api.qualityclouds.ai/mcp"
    }
  }
}

README

[Norma by Quality Clouds]

Norma by Quality Clouds: MCP Server

Norma by Quality Clouds checks AI-written code against your rules as it is written, and keeps a record of what it found.

Norma gives your coding agent the rules for your detected stack before it writes, checks each file against them, and records the outcome. The same file and the same rules return the same verdict every time, so what you get back is a record you can show someone rather than an opinion that changes between runs.

  • Remote server (Streamable HTTP): https://api.qualityclouds.ai/mcp
  • Auth: OAuth. Sign in via browser on first connection. Permanent free tier.
  • Requires a linked repository: live_check only runs on a repository linked in your Norma account (GitHub or Bitbucket). Your code is never stored there
  • Works with Cursor, Claude Code, Lovable, Replit and any MCP-compatible client
  • Stacks covered: TypeScript, JavaScript, Python, PHP, Dart and Flutter, Node, React, Vite, Supabase, FastAPI, SQLAlchemy, Adobe Magento and more

What it does

Rulesets are stack-specific and activate automatically for the stack the server detects. A single repository scan activates rulesets across six functional areas: Security, Performance, Scalability, Manageability, Maintainability and Architecture. The Supabase ruleset alone covers 12 rules, including no service-role keys outside the server, no hardcoded keys or project URLs, mandatory error checks on every mutation, no client-side JWT decoding, and rate limiting on edge functions.

The server exposes five tools:

ToolWhat it does
get_rulesetsDetects your repository's tech stack, with no configuration, and returns the applicable rulesets, one per language or framework
get_rules_for_rulesetReturns the rules in one ruleset: each rule's ID, name, severity, impact area, description and fix guidance
live_checkDeterministic check of one file against your organization's rules. Same file, same rules, same verdict every time. Nothing is stored. Runs only on a repository linked in your Norma account
get_open_issuesReturns the open issues from the last full scan of the linked repository, each with the context needed to fix it
register_applied_actionsRecords what was done in your compliance audit trail: rules verified compliant, violations fixed (file and lines), violations prevented during generation, and which model did the work

Every session produces a structured record of what was checked, fixed and prevented. Your repository's Production-Ready Score and full findings live in your workspace at norma.qualityclouds.com.

How it works

Before the first check, link your repository in your Norma account (see the 5-minute guide below). After that, a coding agent connected to Norma follows this sequence:

  1. get_rulesets: at the start of any coding task.
  2. get_rules_for_ruleset: for each relevant ruleset ID; never skipped.
  3. The agent writes or modifies code, with those rules in context.
  4. live_check: after each file is created or modified, before moving on.
  5. register_applied_actions: after the task, using the exact rule IDs from step 2.

To work through standing issues instead of writing new code, the agent calls get_open_issues and fixes the violations from the repository's last full scan.

Try it in 5 minutes (Claude Code)

  1. Link your repository in Norma. Sign in at norma.qualityclouds.com (free to start, no credit card) and link the repository through the GitHub or Bitbucket integration. Live checks only run on a linked repository. Your code is never stored in your Norma account: linking tells Norma which repository a check belongs to, and live_check holds file content in memory only for the duration of the check.

  2. Add the server:

bash
claude mcp add --scope user --transport http norma https://api.qualityclouds.ai/mcp
  1. Check it connected:
bash
claude mcp list
  1. On first use, your browser opens to sign in.
  2. Open a session in your local clone of the repository you linked and ask Claude Code to review a file against your coding standards. It will fetch the rules for your stack, live check the file, and register the results back to your workspace.

If a live check says the repository isn't linked, confirm it appears in your Norma account and that git remote get-url origin in your clone points at that same repository.

Cursor

Add to .cursor/mcp.json:

json
{  "mcpServers": {    "norma": {      "url": "https://api.qualityclouds.ai/mcp"    }  }}

Cursor will prompt you to sign in via browser on first connection. As with Claude Code, link the repository in your Norma account first (step 1 of the 5-minute guide), then open your local clone of it.

Lovable

Lovable supports custom MCP servers as chat connectors, so Norma works with no plugin. In your project, open the connectors panel and add a custom MCP server named Norma with server URL https://api.qualityclouds.ai/mcp (HTTP transport, OAuth: your browser opens to sign in on first connection). Then send one message, and Lovable keeps it as a standing rule for the project:

From now on, for every file you create or change, run it through Norma and fix the issues it returns.

One step that is not optional: link the project's GitHub repository in your Norma account, then paste its public GitHub URL in chat so Lovable can tell Norma which repository it is working on. Lovable's own git remote is a tokenised URL it will not send to a third party, so without the public URL live checks do not run and nothing reaches your audit trail. Your code is never stored in your Norma account.

Full guide, including troubleshooting: docs/LOVABLE.md.

Links

Built by Quality Clouds, the AI Code Governance platform, governing 950+ enterprise platform instances since 2017.

來源:README.md,提交 991a6ae

工具

0
該伺服器需要登入,因此 SourceWeft 無法在沒有帳號的情況下讀取其工具,也未收錄。已連接帳號所能看到的內容可能因帳號及其權限而異。

版本歷史

2
  1. v1.0.1最新Oct 8, 2026
  2. v1.0.0Sep 16, 2026

評分

已滿足適用檢查項的 85%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具15 分

    請讓伺服器在無憑證時回應 tools/list,或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分