# qa-agent

> An automated, agentic exploratory testing tool that performs comprehensive QA testing on web applications, simulating human user interactions through various input methods (mouse, keyboard, TAB naviga

- **URL**: https://www.freshcrate.ai/projects/qa-agent
- **Author**: billrichards
- **Category**: Testing
- **Latest version**: `v0.2.3` (2026-05-22)
- **License**: MIT
- **Source**: https://github.com/billrichards/qa-agent
- **Language**: Python
- **GitHub**: 1 stars
- **Registry**: github
- **Tags**: `agentic`, `agentic-qa`, `playwright`, `playwright-python`, `python`, `qa`, `qa-automation`, `qa-automation-test`, `qaautomation`

## Description

An automated, agentic exploratory testing tool that performs comprehensive QA testing on web applications, simulating human user interactions through various input methods (mouse, keyboard, TAB navigation, arrow keys, etc.).

## Recent releases

| Version | Date | Urgency | Changes |
| --- | --- | --- | --- |
| `v0.2.3` | 2026-05-22 | High | ### Fixed  - **CSP / stuck-queued bug**: All inline `<script>` blocks extracted from Jinja2 templates to external static files (`static/index.js`, `run.js`, `sessions.js`, `session.js`), resolving `script-src 'self'` violations that blocked SSE event handling and caused the web UI to show "Queued" indefinitely. - **Stop button**: Now correctly reflects actual job state on page load via a pre-flight `/api/status` check before opening the SSE stream. Button text updated to "Stopping after current |
| `v0.2.2` | 2026-05-02 | High | ### Added  - `CONTRIBUTING.md` with development setup, architecture overview, and contribution guidelines. - `docs/README.md` as a documentation hub linking to all detailed docs. - `docs/development.md` with detailed development guide covering environment setup, testing, building, and debugging. - `docs/api-reference.md` with complete programmatic API reference for embedding QA Agent in Python code.  ### Changed  - README.md trimmed for PyPI: Architecture, Development, Contributing, and Program |
| `v0.2.1` | 2026-04-14 | High | Latest release: v0.2.1 |
| `v0.2.0` | 2026-04-14 | Medium | ### Added  - **Multi-LLM support**: Anthropic (Claude) and OpenAI (GPT-4o) providers via Python's built-in `urllib` — no third-party AI SDK required. New `--llm` and `--ai-model` CLI flags. New `llm_client.py` module. - **AI plan validation warnings**: `validate_plan()` detects four patterns in LLM-generated test plans — steps with no assertions, unknown assertion types, hover actions without wait, and trivially-passing hover+visible assertions. Warnings appear in console, Markdown, and JSON re |
| `v0.1.1` | 2026-04-10 | High | ### Added  - Comprehensive test suite: unit tests for `agent`, `cli`, `config`, `models`, `plan_cache`, all reporters, and all testers; web server tests; integration smoke tests against local HTML fixtures. (~3 400 lines across 12 test files, PR #8). - GitHub Actions CI workflow (`test.yml`) running the full test suite on push and pull request. - HTML test fixtures (`tests/fixtures/site/`) used by integration and unit tests. - `tests/conftest.py` shared fixtures (mock Playwright page, mock sess |
| `v0.1.0` | 2026-04-07 | Medium | ### Added - Initial public release. - Agentic exploratory QA testing powered by Claude (Anthropic API). - Two test modes: `focused` (given URLs only) and `explore` (crawl and discover). - Five standard test suites: keyboard navigation, mouse interaction, form handling, accessibility (WCAG), and error detection. - Opt-in WCAG 2.1 AA compliance tester (`--wcag-compliance`): covers non-text contrast, use of color, content on hover/focus, meaningful sequence, input purpose, focus visible, label in |

## Citation

- HTML: https://www.freshcrate.ai/projects/qa-agent
- Markdown: https://www.freshcrate.ai/projects/qa-agent.md
- Dependencies JSON: https://www.freshcrate.ai/api/projects/qa-agent/deps

_Generated by freshcrate.ai. Indexes github releases for AI-agent ecosystem packages._