← Overview
GitHub Trending · Sunday

15 November 2026

15 repos rising that day. Updated

AI agentplan or diffPlannotatorlocal browserfeedbackone clickInboxdecisions, plans
+408 that day★ 11.2KTypeScript

Local review space to annotate agent plans and diffs

Plannotator is a local, browser-based review surface for coding agents. You annotate plans or diffs and send the feedback back to the agent in one click.

  • Works with Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, DeepSeek Harness, and Pi.
  • Besides plans and specs, it annotates markdown and interactive HTML, and reviews diffs and PRs.
  • Plannotator Inbox surfaces decisions, prototypes, variations, and plans from multi-agent work in one place.
source repoauthor, docsskillsweb practicescoding agentmodern codecalibrateand evaluate
+1.2K that day★ 6.1KTypeScript

Modern web guidance for coding agents, from the browser teams

This repo is the source of Modern Web Guidance, a set of skills that bring web platform expertise and browser compatibility data to coding agents. It holds the code and docs for authoring, calibrating, and evaluating that content.

  • It steers coding agents away from legacy patterns and toward modern web platform capabilities.
  • Stewarded by the Google Chrome, Microsoft Edge, Mozilla Firefox, and MDN teams with the developer community.
  • An Eval Dashboard shows the evaluation results for the guidance content.
investortime-limitedskillsClaude workflowsdecisionsyour ownchartingtechnicalcalendarseconomicscreeners
+553 that day★ 4.9KPython

Claude Code skills for investors: market analysis and trading strategy

Claude Trading Skills is a Claude Skills-based workflow toolkit for time-constrained individual investors. It covers market analysis, technical charting, economic calendars, screeners, and strategy development.

  • It aims to structure market review, risk management, trade planning, journaling, and continuous improvement, not to hand buy or sell decisions to AI.
  • Designed for long-term investors with ETFs and dividend stocks at the core, plus disciplined swing trading when conditions favor it.
  • The repo states it is not a signal service and does not promise profitability.
coding agentClaude, Cursordevtools-mcpMCP or CLIlive Chromeinspect, controltracesperformancedebuggingnetwork, consolepuppeteerautomation
+664 that day★ 56.3KTypeScript

Chrome DevTools for coding agents, as an MCP server or CLI

chrome-devtools-mcp lets a coding agent control and inspect a live Chrome browser. It is available as an MCP server and as a CLI.

  • Records traces with Chrome DevTools to extract actionable performance insights.
  • Advanced browser debugging: network request analysis, screenshots, and console messages with source-mapped stack traces.
  • Automation uses puppeteer. Agents such as Antigravity, Claude, Cursor, or Copilot can use it.
git diffchanged filespipelinedeterministicLLM agentreads, searchesocr scanwhole filescommentsline-level
+893 that day★ 51.5KGo

AI code review with deterministic pipelines and an LLM agent

Open Code Review is a CLI for AI code review that began as Alibaba's internal assistant. It reads Git diffs and an LLM agent writes line-level comments.

  • The agent can read full files, search the codebase, and inspect other changed files for context.
  • ocr scan reviews whole files to audit a codebase with no meaningful diff. A built-in ruleset covers NPE, thread-safety, XSS, and SQL injection.
  • Configure an OpenAI or Anthropic compatible model endpoint to get started.
terminalor editorQwen CodeAI coding agentyour projectcode tasksdesktopbrowser, chatnpm packageqwen-code
+64 that day★ 30.4KTypeScript

Open-source coding agent for terminal, editor, desktop, browser, and chat

Qwen Code is an open-source AI coding agent that lives in your terminal. Per its README, it is also available in editor, desktop, browser, and chat.

  • Installed as the npm package @qwen-code/qwen-code.
  • The README marks the Qwen Code 1.0 Preview release and is available in many languages.
your appor AI agentFacade APIconnectsUniver coreSheets, Docseditingcalculationcollaboration
+495 that day★ 24.5KTypeScript

Office SDK for AI: open-source Sheets and Docs for your product

Univer Office SDK brings editing, calculation, collaboration, and file processing into your product for people and AI agents. Built on an extensible open-source core.

  • Start with open-source Sheets and Docs, then connect your own application or agent.
  • The README has an open-source spreadsheet editor quick start and an architecture guide, including feature boundaries and cross-feature edits with undo/redo.
  • Applications and agents connect through the Facade API.
AI agentbackendAG-UIevent streamfrontend appuser contextnpxcreate-ag-ui-app
+9 that day★ 17.4KTypeScript

Event-based protocol connecting AI agents to frontend applications

AG-UI is an open, lightweight protocol that standardizes how AI agents connect to user-facing applications. Event-based, it integrates agents, real-time user context, and UIs.

  • A new AG-UI app is created with a single npx command.
  • There is an AG-UI Dojo, quickstarts for building apps and new framework integrations, and a Discord community.
  • The core package is on npm as @ag-ui/core.
trynpx create-ag-ui-app my-agent-app
checkpointmodel filesZig binarytokenize, serveGPU kernelsMetal, CUDArequesttext inVulkanalso listed
+37 that day★ 2.2KZig

LLM inference engine in one Zig binary for Metal, CUDA, and Vulkan

TensorFold serves language models from a Zig binary on Apple Silicon and NVIDIA GPUs. The engine reads checkpoints, tokenizes requests, and runs Metal or CUDA kernels directly.

  • Serving needs no Python or MLX installation.
  • The repo description lists Metal, CUDA, and Vulkan support.
  • Contributors and agents are asked to use Agent Traffic Control (ATC) before editing to avoid duplicate work.
coding agentwordy outputcavemanskill + proxyfewer tokensshort repliesCaveman Codenew, free agent
+227 that day★ 120KGo

Skill and proxy that make coding agents talk like a caveman

Caveman is a skill and proxy for coding agents that, per its description, cuts 65% of tokens by talking like a caveman. Code stays exact.

  • The README cites 33.2% fewer input tokens over whole sessions, from 54 Claude Code runs with 18 of 18 answers right.
  • A new free product, Caveman Code, is a coding agent claimed to be 14% cheaper than Claude Code on Terminal-Bench 4, from the authors' own runs.
  • These numbers come from the repo authors' own measurements, not independent ones.
monitorsHTTP, TCP, pingUptime Kumaself-hostedalerts90+ servicesDockercontainers20 scheck interval
+430 that day★ 94.3KJavaScript

Self-hosted uptime monitoring with a fast, reactive UI

Uptime Kuma is an easy-to-use self-hosted monitoring tool. It monitors uptime over many protocols and sends notifications to many services.

  • Monitors HTTP(s), TCP, HTTP(s) keyword, HTTP(s) JSON query, Websocket, Ping, DNS, Push, Steam Game Server, and Docker containers.
  • Notifications via Telegram, Discord, Gotify, Slack, Pushover, Email (SMTP), and 90+ other services. Check interval is 20 seconds.
  • The README warns that uptimekuma.co is not the official site and recommends following the install guide in the repo.
Maybearchived repoSurecommunity forkfinance appfor everyonev0.6.0final release
+659 that day★ 15.3KRuby

Personal finance app, a community fork of Maybe Finance

Sure is a personal finance app for everyone, written in Ruby. This repo is a community fork of the now-abandoned Maybe Finance project.

  • The Maybe Finance team spent most of 2021 to 2022 building a full-featured wealth management app.
  • Maybe once had an Ask an Advisor feature connecting users with a CFP/CFA, included in the subscription.
  • The README links Maybe's final release, v0.6.0, for more detail.
Main Botyou chat a needDocker sandboxchecks out repospull requestopenedSOPsaved workflowproject rolesdelegated workhuman gateapproval
+33 that day★ 934Go

Self-hosted workflows for coding-agent delivery

LoopForge is a self-hosted workspace where teams ship software with coding agents. You describe a need to Main Bot in chat, and it opens a Docker sandbox, checks out repos, delegates to project roles, and opens the PR.

  • A way of working that proves itself can be saved as an SOP, a standard workflow of agents, scripts, and human approval gates.
  • Agent CLIs used include CodeBuddy, Codex, Claude Code, OpenCode, Cursor CLI, Antigravity, Kimi Code, Pi, DeepSeek Harness, and ZCode.
  • Runs on your own infrastructure under the MIT license.
sourcevideo, captureGPU pipelinereal timepreviewenhancedSuper ResupscalingDenoise, HDRFrame Genmore frames
+420 that day★ 2.3KC++

Windows video player with real-time GPU enhancement

Veyra is a Windows video player and capture enhancement tool with real-time preview for videos, images, and capture card sources. It combines AI Super Resolution, DLSS5 enhancement, and Frame Generation.

  • The repo description claims it can bring PS5 GTA6 to 120FPS with DLSS 5. This claim was not verified here.
  • Features include Super Resolution, denoising, HDR, Frame Generation, capture, and PS5, Xbox, or PC streaming.
  • The README cites 79K+ portable package downloads (79,332 as of 2026-11-08). Those are download events, not unique users.
WiFi trafficnearbyAirgorahRust + GTK4audit datahandshakesclientsdiscoveredown networkslegal use only
+270 that day★ 5.9KRust

WiFi security auditing app built with Rust and GTK4

Airgorah is a Linux WiFi security auditing tool that captures nearby WiFi traffic and discovers clients connected to access points. It can also capture handshakes and PMKIDs.

  • Written in Rust, with a GTK4 interface.
  • The README says it is for testing networks you own. Attacking networks you do not own is illegal in almost all countries.
  • Runs on Linux only and is available through crates.io and the AUR.