Qualimetry AI App
Qualimetry/vscode-qualimetry-ai-app
Valida código de IA contra padrões organizacionais antes de review
Ver no GitHubExtensão para VS Code e Cursor que integra seu assistente de IA a um servidor Qualimetry, capturando violações de padrões de código, princípios e políticas durante a autoria. Valida automaticamente o código gerado pela IA antes da revisão, eliminando achados de conformidade em tempo de escrita.
Instalação
npx convez-ai Qualimetry/vscode-qualimetry-ai-app1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Garantir conformidade de código gerado por IA com padrões corporativos
- Detectar violações de segurança e políticas durante autoria, não após review
- Integrar achados de análise estática e resolvidos de CVEs em chat
Veredito

Ferramenta robusta de compliance que fecha um gap real: validação de código IA antes de review. A arquitetura agnóstica (VS Code, Cursor, VSCodium) e setup sem JSON a tornam acessível, mas depende de infraestrutura Qualimetry própria, limitando adoção para quem não usa a plataforma.
README
A tradução para o português deste README ainda está na fila.
Qualimetry AI App for VS Code & Cursor
Catches policy violations before code review, not during. The Qualimetry AI App keeps every line of code your AI assistant writes — and every reviewed file you touch — aligned with your organisation's coding standards, principles, and policies, automatically.
The same .vsix artefact is published to:
- Visual Studio Marketplace — for VS Code Stable and Insiders.
- Open VSX Registry — for Cursor, VSCodium, Theia, Eclipse Che, Gitpod, and any other VS Code-compatible editor.
What it does
- Connects your editor's AI agent to your Qualimetry server. Once configured, Copilot Chat (in VS Code) and Cursor's chat agent can pull your organisation's coding standards, general coding principles, secure coding principles, and policies on demand for any file you're working on.
- Keeps the AI's own code compliant from the first line. Ask the agent to write or modify code and it silently fetches the four pillars for the file's language and applies them. High-severity violations are corrected before the code is shown to you.
- Pulls open review findings on request. Ask the agent "show me the qualimetry findings for this file" — it returns the violations grouped by pillar (Coding Standards, Design & Best Practice, General Principles, Secure Principles, Policies) and severity, plus a standards-compliant example of how to resolve them. Ask for a specific pull request's findings to scope the results to the issues raised on the pull request's new code.
- Triages rules-based analysis findings. Bugs, security vulnerabilities, and code smells from your static analysis (Sonar-style severities) are accessible from chat. Supply a pull-request number to target the pull request's new-code issues. (Qualimetry Enterprise.)
- Resolves dependency CVEs. The agent walks vulnerable packages and proposes safe upgrades to the next non-vulnerable version. (Qualimetry Enterprise.)
Benefits
- Catch policy + standards violations during authoring and before review, not after.
- AI-written code is compliant by construction — no separate "lint pass" needed.
- One install, one setup, no further configuration per repo. The setup flow uses VS Code's native input boxes — you never touch JSON.
- Same
.vsixworks for VS Code, VS Code Insiders, Cursor, VSCodium, and every Open VSX-compatible editor.
Quick Start - VS Code
-
Open VS Code.
-
Press
Ctrl+Shift+X(Windows / Linux) orCmd+Shift+X(Mac) to open the Extensions view. -
In the search box, type:
Qualimetry AI App -
Click Install on the Qualimetry AI App extension.
-
Look for a notification in the bottom-right corner of the window that says:
Qualimetry: configure to get started [Configure]Click the blue Configure button.
-
A text box will appear at the top of the window. Type your Qualimetry server URL (or accept the default
https://myorg.qualimetry.io/mcp/by pressing Enter without typing). -
A second text box will appear, this one for your access token. The text you paste will be hidden. Paste your token and press Enter.
-
You will see a green notification confirming setup, with a button Install Skills. Click it once to enable the bundled agent workflows.
That's it. VS Code does not need to be restarted; the Qualimetry server is picked up automatically.
If you missed the original notification, run setup any time:
- Press
Ctrl+Shift+P(Windows / Linux) orCmd+Shift+P(Mac) to open the Command Palette. - Type:
Qualimetry: Setup - Press Enter on the matching entry.
- Follow the same two prompts above.
Quick Start - Cursor
-
Open Cursor.
-
Open the Cursor Marketplace (search for "qualimetry"), or — for manual install — paste a
qualimetryblock referencing your server URL and access token into~/.cursor/mcp.json. A ready-made template lives in this repo atvscode-client/snippets/cursor-mcp.snippet.json. -
After install, click the Configure button on the toast notification, or run Command Palette → Qualimetry: Setup (the extension is the same
.vsixas the VS Code build, published to Open VSX which Cursor uses). The extension detects Cursor and writes~/.cursor/mcp.jsonfor you. -
Restart Cursor for the configuration to load.
Configuration
The extension exposes two settings in File → Preferences → Settings → Qualimetry:
| Setting | Default | What it controls |
|---|---|---|
qualimetry.mcpServerUrl | (empty) | Your Qualimetry server URL. The Setup command writes this for you. The token is never stored in settings. |
qualimetry.scope | user | Where setup registers the server. user: VS Code's extension-provided MCP registry (token in VS Code secret storage), or ~/.cursor/mcp.json in Cursor. workspace: <repo>/.vscode/mcp.json (token prompted via VS Code's inputs mechanism, never written to the file), or <repo>/.cursor/mcp.json in Cursor. |
Where credentials live
After setup:
- VS Code (user scope, default): the server is provided to VS Code by the extension itself (MCP server definition provider). The URL lives in the
qualimetry.mcpServerUrlsetting; the access token lives in VS Code's encrypted secret storage — no JSON file holds it. - VS Code (workspace scope):
<repo>/.vscode/mcp.jsonholds the server entry with the token referenced via aninputsprompt, so the secret never lands in a file that might be committed. - Cursor:
~/.cursor/mcp.json(user scope) or<repo>/.cursor/mcp.json(workspace scope) — Cursor's native MCP store.
Limitation: explicit, not proactive
VS Code Copilot and Cursor do not currently expose a tool-event hook the way Claude Code and Codex do. As a result, this extension cannot automatically surface review findings the moment a reviewed file is opened. Use /review-check <file> in Copilot Chat / Cursor chat instead, or ask the agent to "use the qualimetry review-check skill on the current file."
Building from source
cd vscode-client
npm install
npm run bundle
npx vsce package --allow-missing-repository
This produces qualimetry-vscode-ai-app-X.Y.Z.vsix in vscode-client/.
License
Apache 2.0. See LICENSE.
Built on the Qualimetry AI Skills workflow library.