Skip to content

Releases: OTweihan/Argus

Release list

v0.1.1 — CI Deprecation Warning Cleanup

v0.1.1 — CI Deprecation Warning Cleanup Latest
Latest

Choose a tag to compare

View all tags
@OTweihan OTweihan released this 22 May 05:28
v0.1.1
e84c42f

Highlights

This release eliminates all DeprecationWarning noise from CI logs and upgrades GitHub Actions to Node 24.

Changes

CI & Actions

  • Upgraded 7 GitHub Actions to latest versions (Node 20 → 24): checkout@v5, setup-uv@v8, setup-node@v6, pnpm/action-setup@v6, cache@v5, setup-buildx-action@v4, build-push-action@v7
  • All builds now use Node 24 runtime

Deprecation Fixes

  • Starlette HTTP status codes: Renamed HTTP_413_REQUEST_ENTITY_TOO_LARGE → HTTP_413_CONTENT_TOO_LARGE and HTTP_422_UNPROCESSABLE_ENTITY → HTTP_422_UNPROCESSABLE_CONTENT per Starlette 1.0 naming
  • TaskService facade removed from tests: All test files now inject TaskLifecycleService and TaskReadService directly instead of going through the deprecated TaskService wrapper
  • TaskApplicationService refactored to accept sub-services directly (lifecycle + task_read)

Full Changelog

  • 16 files changed, 344 insertions(+), 257 deletions(-)
  • Zero new warnings in strict mode (-W error::DeprecationWarning)
  • All 563 unit + integration tests passing
Assets 2
Loading

v0.1.0 — AI Native Test Platform

v0.1.0 — AI Native Test Platform

Choose a tag to compare

View all tags
@OTweihan OTweihan released this 21 May 07:05
v0.1.0
5adde70

Argus initial public release.

Overview

Argus is an AI-native test platform for Web application QA. Describe test goals in natural language, and Argus handles planning, browser automation, and reporting.

Features

  • Natural language test execution — argus run --goal "..." --url "..." drives black-box testing
  • LLM-driven Planner & Evaluator — Two specialized prompts with business-rule extension support
  • Playwright browser automation — Chromium headless/headed, screenshots, DOM snapshots
  • Self-healing execution — Failed actions trigger recovery (2 retries by default) instead of aborting
  • Browser auth state management — argus auth save/list — login once, reuse across tasks
  • Structured reporting — HTML + JSON dual format with screenshots and issue tracking
  • Task observability — SQLite timeline, real-time WebSocket events, LLM call traces, debug bundles
  • Web Console — Vue 3 + Element Plus SPA for managing projects, tasks, and models
  • REST API + WebSocket — FastAPI backend with OpenAPI documentation
  • Prompt extension system — Per-project and per-task business rule customization
  • Sensitive data redaction — Automatic masking of API keys, passwords, tokens in logs and traces
  • Docker deployment — Containerized with SSRF/CORS/rate limiting/token auth

CLI Commands

argus serve             Start the web server
argus run               Execute a black-box test
argus browser check     Debug browser capabilities
argus auth save/list    Manage browser login states
argus llm check         Verify LLM connectivity
argus config llm        Configure the LLM

Tech Stack

  • Python 3.11+ / FastAPI / Playwright
  • Vue 3 + TypeScript + Element Plus
  • SQLite / Jinja2 / Docker
Loading