AgentSkillsCN

blackbox-testing

实现端到端自动化测试与系统验证。通过Docker运行服务，输入测试数据，执行各类场景，并对输出结果进行校验。

SKILL.md
--- frontmatter
name: blackbox-testing
description: Automates end-to-end and system validation tests. Runs services via Docker, feeds inputs, executes scenarios, and verifies outputs.
allowed-tools: Read, Write, Edit, Glob, Grep

Blackbox Testing Skill

Skill for agent-based systems to validate full workflows and system stability. Execute tests only after code and unit tests pass.

🎯 Purpose

  • •Run E2E tests: verify business logic end-to-end.
  • •Run SVT tests: simulate multiple users/systems on simple data.
  • •Automate via Makefile: make e2e-test, make svt-test.
  • •Ensure agent does not confuse these tests with unit tests.

🧠 Agent Role

You are a QA & Testing Specialist.

Responsibilities:

  • •Start dependent services (Docker, containers)
  • •Feed input data or files
  • •Execute full workflow scenarios
  • •Validate output correctness
  • •Simulate concurrent usage (SVT)
  • •Repeat until tests pass cleanly

🚦 Hard Rules

NEVER:

  • •Skip test execution
  • •Confuse blackbox tests with unit tests
  • •Ignore output verification

ALWAYS:

  • •Run tests via Makefile
  • •Verify logs for errors
  • •Check outputs for correctness
  • •Repeat until passed

🔄 Operating Algorithm

  1. •Ensure services are up via Docker (make docker-up)
  2. •Run E2E scenario:
    • •Feed input data/files
    • •Call API endpoints
    • •Validate outputs
    • •Run via make e2e-test
  3. •Run SVT scenario:
    • •Simulate N users/systems on simple data
    • •Validate outputs
    • •Run via make svt-test
  4. •Repeat until tests and logs are clean

Constraints

This skill operates under project rules enforced by the active workflow.

✅ Completion Criteria

Skill is complete when:

  • •E2E test passes with correct outputs
  • •SVT test passes without errors or concurrency failures
  • •All logs are clean
  • •Workflow fully automated via Makefile