Antigravity Testing Kit: An End-to-End AI Agent Workflow Suite for the Software Testing Domain

This post introduces Antigravity Testing Kit, a comprehensive AI Agent configuration suite designed for the software testing community. It supports the full lifecycle of both manual and automated testing, integrating AI capabilities into every stage (requirements analysis, test case design, execution, reporting). Key features include modular architecture (rules, skills, workflows), compatibility with mainstream tools (Playwright, Selenium, Appium), and seamless integration with R&D toolchains (Jira, Xray). This suite aims to help testing teams improve efficiency while maintaining quality, bridging AI and traditional testing processes.

With AI rapidly permeating software development, testing engineers face challenges in integrating AI into existing workflows while balancing quality and efficiency. Developed by Vietnamese testing community contributor Anh Tester, Antigravity Testing Kit addresses this need. Unlike tools focusing only on automated script generation, it covers the full testing lifecycle and supports both manual and automated modes—an uncommon feature in similar projects. Its end-to-end design philosophy ensures AI assistance across all testing stages.

The kit uses a modular architecture with three core components under .agent/:

1. Rules System: Located in .agent/rules/, it defines AI Agent behavior guidelines (e.g., Page Object Model, smart waiting, locator repair) to ensure output quality and consistency.
2. Skills System: 10 specialized modules (automation engineer, manual testing, UI debugging, locator repair, test data generation, framework architect, Jira integration, etc.) enable flexible AI capability allocation.
3. Workflow System: 15 predefined slash commands (e.g., /generate_automation_from_testcases, /generate_manual_testcases_rbt) break complex tasks into repeatable steps, reducing cognitive load for AI collaboration.

Manual Testing Support: The kit provides AI-assisted risk-based test design (AI-RBT), which analyzes requirements to identify risk areas and generate prioritized test cases. Its structured workflow (requirement parsing → risk identification → test design → test case generation → review and optimization → report output) ensures complete coverage while saving time. Automated Testing Support: Covers the full chain from framework initialization (Playwright TypeScript/Selenium Java skeletons) to script generation (natural language to executable code), code review (static analysis), flaky test analysis, and API test generation (from Swagger docs, supporting Playwright/REST Assured).

External Tool Integration: Seamlessly integrates with Jira/Xray (requirement synchronization, test result reporting) and Google Sheets (test data import/export). Multi-Platform Support: Web testing (Playwright/Selenium), mobile (Appium), API (Playwright/REST Assured). This flexibility allows teams to use existing tech stacks without reconstruction.

Usage Modes:

• Complex Task Mode: Follow plans/ guides for multi-step tasks (e.g., framework setup).
• Quick Task Mode: Use prompt_templates/ for single tasks (e.g., code generation).
• Workflow Mode: Trigger slash commands for standardized tasks. Target Scenarios:
• Teams transitioning to AI-assisted testing.
• Multi-tech-stack environments.
• Agile teams needing fast regression testing.
• Organizations building/optimizing QA systems.

Current limitations include:

1. Documentation is mainly in Vietnamese, creating barriers for non-Vietnamese users.
2. Long-term maintenance depends on community contributions, with sustainability to be observed.
3. It's exclusive to Antigravity platform, limiting its universality. Improvement areas: Localize documentation to more languages, strengthen community building for sustained development, and explore cross-platform adaptability.

Antigravity Testing Kit combines domain expertise, prompt engineering, and workflow orchestration to provide a practical AI collaboration solution for testing. It represents a new paradigm where humans handle strategy/decision-making and AI handles execution, balancing efficiency and quality. As AI evolves, domain-specific AI Agent suites like this are expected to become industry standards, and this kit offers valuable references for the trend.