convez.ai
AgenteAutomaçãoIntermediário
Piyabordee

Test Copilot — Agent de Engenharia de Testes

Piyabordee/test-copilot

Gera estratégias e código de teste com aprovação gated para qualquer projeto

Ver no GitHub
InicialPluginScore · de 0 a 100

Agent IA que automatiza a criação de estratégias de teste, casos de teste e checklists de segurança para APIs REST, GraphQL, aplicações web, mobile e microsserviços. Implementa workflow AI-DLC com aprovação humana antes da geração de código.

Instalação

$npx convez-ai Piyabordee/test-copilot

1 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

Estrelas
1
Forks
0
Acompanhando
0
Saúde
28%
Último commit
há 5 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Criar estratégias de teste abrangentes para APIs REST e GraphQL
  • Gerar casos de teste e código executável em frameworks existentes
  • Validar cobertura de segurança (OWASP, autenticação, autorização)
  • Testar aplicações web, mobile e microsserviços com especialização
  • Propor configuração de CI pipeline condicional após aprovação

Veredito

Agent maduro com workflow bem estruturado (aprovação gated antes de escrita). Cobre amplitude impressionante de tipos de projeto, mas exige disciplina humana na Step 1 para evitar código ruim na Step 2 — ideal para times que valorizam validação antes de execução.

README

A tradução para o português deste README ainda está na fila.

Test Copilot

AI-powered test engineering agent for creating, reviewing, and improving test strategies and test cases for any type of software project.

It follows an AI-DLC workflow — Generate → Human Validate → Execute — split into two approval-gated steps so AI drafts and humans approve before any code is written.

Installation

Via Claude Code Marketplace (Recommended)

In Claude Code, type:

GitHub: Piyabordee/test-copilot

The plugin will be automatically discovered and installed.

Manual Installation

See INSTALLATION.md for step-by-step manual setup.

How to Use

Step 1 — Generate a test strategy (read-only):

/test-copilot:test-engineer

The agent analyzes the codebase, produces a complete test strategy, then stops and asks you to review. Iterate until you are satisfied, then reply APPROVE.

Step 2 — Turn the approved strategy into code (writes files):

/test-copilot:generate-tests

The agent generates runnable tests that mirror the approved cases, using your existing test framework, and conditionally proposes a CI pipeline.

Command Overview

/test-copilot:test-engineer — Strategy (Step 1 of 2)

Analyze the codebase and produce a test strategy and test cases for human review. Does not write files.

Execution flow:

  1. Gathers business context, detects project type & test framework, and inventories existing tests
  2. Invokes the test-engineer-skill for deep testing knowledge
  3. Produces a complete test strategy and self-checks it
  4. Stops and asks for approval — reply APPROVE when satisfied

/test-copilot:generate-tests — Code (Step 2 of 2)

Turn the approved strategy into executable test code and CI config. Writes files. Requires an approved strategy in the conversation.

Specializations:

  • REST API Testing (contract testing, rate limiting, security headers)
  • GraphQL Testing (query complexity, nested resolvers, schema integrity)
  • Mobile Application Testing (network conditions, device fragmentation, lifecycle)
  • Web Application Testing (cross-browser, responsive design, accessibility)
  • Microservices Testing (circuit breakers, service discovery, distributed tracing)
  • Security Testing (OWASP Top 10 injection, auth, authorization, data exposure)
  • Performance Testing (response time percentiles, throughput, load testing)

Architecture

Step 1 — Strategy (read-only)          Step 2 — Code (writes files)
/test-copilot:test-engineer            /test-copilot:generate-tests
   │                                       │
   └── test-engineer-agent                 └── test-codegen-agent
        ├── Reconnaissance & Context             ├── Load approved strategy
        ├── Invoke Skill                         ├── Invoke Skill (codegen)
        ├── Generate Strategy                    ├── Generate code (existing framework)
        ├── Self-Check                           ├── Place files (ai-generated/)
        └── STOP → reply APPROVE ───────────▶    ├── CI & secrets (conditional)
                                                 └── Self-Check
                         │                               │
                         └────── test-engineer-skill ────┘
                          (knowledge, templates, checklists, codegen rules)

Example

Step 1 — in a Node.js REST API project, /test-copilot:test-engineer produces:

# Test Strategy: My API Project

## Detected Project Type
- Primary: REST API (Express.js)
- Test Framework: Jest + Supertest

## Test Cases

### Suite: Auth API (`/api/auth`)
| # | Test Case | Method | Endpoint | Input | Expected | Status |
|---|-----------|--------|----------|-------|----------|--------|
| 1 | Login with valid credentials | POST | /api/auth/login | {email, password} | 200 + JWT | Critical |
| 2 | Login with wrong password | POST | /api/auth/login | {email, "wrong"} | 401 + error | Critical |
| 3 | Rate limit after 5 failures | POST | /api/auth/login | 6x failed login | 429 + Retry-After | Critical |

### Recommended Next Steps
1. Add test files: tests/auth.test.ts, tests/users.test.ts
2. Run: npm test -- --coverage

> Please review this test strategy. Reply APPROVE, then run /test-copilot:generate-tests.

Step 2 — after you reply APPROVE, /test-copilot:generate-tests writes the actual tests/*.test.ts files using Jest + Supertest and proposes a CI step.

Ver o original no GitHub

Bom para

Relacionados