20 Claude Code testers run mobile QA on GitHub runners
An open-source repo runs 20 Claude Code agents on iOS Simulators and Android emulators at once, hunting broken flows, layout and translation bugs for about $5 a run.
How coding agents work and where they help.
An open-source repo runs 20 Claude Code agents on iOS Simulators and Android emulators at once, hunting broken flows, layout and translation bugs for about $5 a run.
A one-script Mac tool gives each Claude account its own window and sign-in, while every account shares a single Claude Code setup and sessions can message each other.
The w01fb0ss/PiBot repository gathered 301 stars in its first day: a desktop app where named AI bots plan, run commands and edit files behind an approval gate.
JimLiu/open-artifact hit 304 stars in five days. It gives coding agents eight artifact types, saved as local files, that you can edit, comment on and export.
A new MIT-licensed repo profiles Python, Node, Ruby and other apps, ranks the slowest functions and has a coding agent port just those to Rust, with replay checks before enabling.
AWS's Agent Toolkit for AWS now includes a distributed KV caching skill that lets coding agents size, set up and validate an ElastiCache cache for self-hosted LLM inference.
Browser Use ships 1.0.8rc1, a pre-release that changes how the judge turn is billed, lets headless and SSH users approve runs from another device, and stops bu.type printing password values.
OpenAI published Agents SDK v1.0.0-alpha.1 on 20 November, adding Codex as an opt-in model provider for the Python SDK, with wheels on GitHub Releases only.
Claude Code v2.1.330 brings enterprise sign-in for organizations managed in Anthropic's Admin Console, a built-in chart mod called easel, and new self-hosted runner options.
browser-use 1.0.7 makes long MCP browser_exec calls reply within 50 seconds with partial output, so clients with a 60 second limit get a result instead of a timeout.
Google's walkthrough shows how Gemini Enterprise reaches internal Cloud Run MCP servers using Egress Agent Gateway, Private Service Connect, scoped DNS and dual-header auth.
Google's guide explains how a custom UI calling Gemini Enterprise's StreamAssist API must handle user identity, connector 3-legged OAuth and per-tool agent auth.
phonedriver, a Rust MCP server with 304 GitHub stars in 12 days, lets Claude Code, Codex and Cursor read and tap Android and iOS screens as text, with a guard before payments.
A new MIT-licensed agent skill, antialgorithm, makes models name their most average idea first and then move away from it. It has 553 GitHub stars in six days.
A developer published the rules file and three review prompts he uses after an estimated 4,500 hours with AI agents. The repo has 304 stars in six days.
The morluto/rea repository puts reverse engineering tools behind one MCP server, letting agents such as Claude Code and Cursor inspect binaries, apps and websites.
ECC, an MIT-licensed agent harness for Claude Code and other coding agents, ships 72 agents, 328 skills and memory hooks, and lists 307,367 GitHub stars.
simplexai-labs/linger drew 1,622 stars in ten days for an open-source personal agent that claims a 125B model on a 12 GB GPU, with nightly memory and proactive follow-ups.
Kimi Code CLI v2.7.0 adds a native Amazon Bedrock provider that authenticates through the standard AWS credential chain, so no separate Bedrock API key is needed.
GitLab's walkthrough installs a skill in Claude Code that configures a whole group for Duo Agent Platform, then hands issue work, code review and pipeline fixes to GitLab.
ai-timon/timon is an open-source agentic workstation that splits a job across coder, reviewer and merger agents, running on the CLI subscriptions you already pay for.
GitLab published recommended configurations pairing Claude Code, Codex, Cursor and other agents with Duo, plus a setup skill for review, CI repair and security triage.
Google's Tool Retrieval is now in preview in the Gemini API, letting agents load tool parameter schemas only when needed instead of on every request.
Version 2.1.328 turns on auto mode for Opus 4.6 and Sonnet 4.6 across Bedrock, Google Cloud's Agent Platform, Foundry and the Claude apps gateway, alongside hooks, runner and session fixes.
GitHub CLI v2.105.0 adds gh event listen, which waits for pull request events such as reviews and check suites instead of polling in a loop. Both it and Event Streams are in preview.
GitHub added eight MCP connectors to Copilot in public preview, including Supabase, Box, Buildkite, Calendly, ClickUp, Mixpanel, Webflow and Wix.
GitHub Copilot team customizations are generally available, letting team maintainers and organization owners set approved Agent Plugins as defaults for direct members.
Google detailed Sashiko, its open-source Gemini-powered agentic code review system for Linux kernel patches: about 1,200 reviews a day and a 35% drop in 60-day fix rates.
Microsoft's Foundry Dev Pack is now generally available, bundling the Foundry CLI, a skill for coding agents, the VS Code toolkit and a GitHub Copilot app plugin in one install.
A TypeScript repo that gathered 301 stars in its first week turns viral short videos into a Viral DNA report, an editable VFML project and 10 costed variants.
Dify 1.20.0 adds OpenTelemetry trace export, malware scanning for knowledge uploads, HITL approvals in email and chat tools, and personal access tokens for its OpenAPI.
opencode v2.3.2 sends the system prompt as a leading developer message on every Responses route so GPT-6.2 caches it, cutting uncached input per step from about 14,000 tokens to 3 on one Copilot model.
AWS has tested three OAuth configurations for AI agents calling Snowflake Cortex AI over MCP, covering per-user sign-in, corporate SSO and shared service accounts.
Vercel's Sandbox Python SDK reaches 1.0, stabilising its API and matching the TypeScript version, with a breaking bytes change for anyone upgrading from 0.x.
GitHub Copilot adds OpenAI's GPT-6.2 Astra and GPT-6.2 Sol for agentic coding and code review, available to Pro+, Max, Business and Enterprise users.
SMART is an open source Python SDK and Terraform stack that deploys a WhatsApp channel layer for agents on Amazon Bedrock AgentCore, covering media, memory, scheduling and Meta templates.
GitHub's Copilot code review now chooses Lite, Balanced or Max per pull request, flags exposed secrets, reads related public repos and ships on Azure DevOps.
GitHub's November 17 Copilot update for JetBrains adds a cloud sandbox, remote access across Copilot clients, browser use and on-device voice dictation, all in public preview.
GitHub's Copilot app now handles Azure DevOps repositories and signed-in browser tasks in public preview, and GitHub Enterprise Server support is generally available.
GitHub expanded Copilot Fusion on November 17, 2026 with three new models, wider surface support, and a 71.2% Terminal-Bench 4.1 score for Fusion Max.
GitHub added MAI Code 2 and MAI Code 2 Flash to Copilot's model picker on November 17, 2026, with Flash reaching Free and Student plans and both billed by usage.
GitHub Copilot Evals is in public preview. It lets you create repeatable tests for skills, custom agents and MCP tools, then compare runs after a change.
GitHub Copilot CLI now supports Microsoft Entra sign-in in public preview, letting EMU enterprises and Microsoft 365 Copilot license holders authenticate without a separate GitHub sign-in.
A day-old GitHub project records what your app does, then checks every piece of a Rust port against that recording. It supports six languages and connects to four coding agents.
Nexus is a TypeScript local AI control plane that connects ChatGPT, DeepSeek and other MCP clients to local files and tools. It reached 362 GitHub stars in four days.
A new MIT-licensed GitHub repo hands coding agents 31 marketing skills covering SEO, landing pages, ads and email, built from the author's own campaigns.
A day-old GitHub repo packages local LLM inference optimization as an agent skill, with benchmarking scripts, ceiling maths and before-and-after numbers on named hardware.
A new managed setting, strictPluginOnlyCustomization, lets GitHub Enterprise Cloud admins require Copilot customizations to come from plugins or managed settings instead of user or project config.