← Overview
GitHub Trending · Monday

16 November 2026

12 repos rising that day. Updated

your appor AI agentFacade API+ eventsUniver coreSheets, DocsOffice SDKCLI, agents
+988 that day★ 26.1KTypeScript

Open-source office SDK: Sheets and Docs for apps and agents

Univer Office SDK brings editing, calculation, collaboration, and file processing into your product for people and AI agents. It is built on an open-source core, starting with Sheets and Docs.

  • The quick start is an open-source spreadsheet editor you can connect to your application.
  • Applications and agents connect through Facade APIs and events, with clear boundaries for custom features.
  • Univer Workspace is included as a reference application for the Office SDK.
AI agentbackendAG-UIevent streamfrontend appuser contextDojoexamples
+355 that day★ 17.9KTypeScript

Open event-based protocol connecting AI agents to apps

AG-UI is an open, lightweight, event-based protocol that standardizes how AI agents connect to user-facing applications.

  • A new app can be created in seconds with npx create-ag-ui-app my-agent-app.
  • The protocol links agents, real-time user context, and user interfaces.
  • The AG-UI Dojo and quickstarts cover building apps and new framework integrations.
trynpx create-ag-ui-app my-agent-app
Stitch skillsAgent SkillsinstallpluginsCodexmarketplaceGemini CLIClaude CodeCursorOpenCode
+247 that day★ 9.3KTypeScript

Agent skills for the Google Stitch MCP server

A collection of agent skills and plugins for Google Stitch, following the open Agent Skills standard. Works with Codex, Antigravity, Gemini CLI, Claude Code, Cursor, and OpenCode.

  • Installing the plugins globally is described as the fastest, recommended route.
  • For Codex, add the marketplace via the CLI, then install the plugins you need.
  • OpenCode is supported through manual install only.
trycodex plugin marketplace add google-labs-code/stitch-skills --ref main
AI agentplan or diffPlannotatorlocal browserfeedbackone clickInboxmulti-agent
+866 that day★ 11.9KTypeScript

Local review space to annotate agent plans and diffs

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

  • Works with Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, DeepSeek Harness, and Pi.
  • It can annotate interactive HTML and review diffs and PRs.
  • Plannotator Inbox surfaces decisions, prototypes, variations, and plans from multi-agent work.
coding agentMCP or CLIdevtools-mcpserverlive ChromeDevToolsoutputtraces, console
+717 that day★ 57KTypeScript

Chrome DevTools for coding agents, via MCP and 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 and extracts performance insights using Chrome DevTools.
  • Analyzes network requests, takes screenshots, and reads console messages with source-mapped stack traces.
  • Used from agents such as Antigravity, Claude, Cursor, or Copilot.
Git diffchanged filesocr + agenttools, rulesLLMLLM endpointcommentsline-level
+691 that day★ 52.2KGo

AI code review CLI: deterministic pipelines plus an LLM agent

Open Code Review is an AI-powered code review CLI that began as Alibaba's internal review assistant. It reads Git diffs and produces line-level review comments through an LLM you configure.

  • The agent can read full file contents, search the codebase, and inspect other changed files for context.
  • ocr scan reviews entire files, not just diffs.
  • Includes a built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection) and is compatible with OpenAI and Anthropic.
you in chatrequirementMain BotDocker sandboxSOPagents + gatesagent CLIsCodex, Claudepull requestyour repos
+423 that day★ 1.6KGo

Self-hosted workflows for coding-agent delivery

LoopForge is a self-hosted workspace where a team ships 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 flow of agents, scripts, and human approval gates.
  • The work is done by coding-agent CLIs such as CodeBuddy, Codex, Claude Code, OpenCode, Cursor CLI, and Antigravity.
  • Roles, sandboxes, checkpoints, and history run on your own infrastructure.
C/C++ codeno dependenciesggmltensor opsbackendsCPU, GPU, NPUtargetsx86, ARM, wasm
+43 that day★ 15.9KC++

Dependency-free C/C++ tensor library for machine learning

ggml is a tensor library for machine learning, written in plain C/C++ with no dependencies, aiming to be simple, portable, and efficient.

  • Runs on x86, ARM, RISC-V, LoongArch, PowerPC, s390x, and WebAssembly.
  • Backends cover CPU, GPU, NPU, and browser, with 2- to 8-bit integer quantization plus MXFP4 and NVFP4.
  • Zero memory allocations during runtime.
trygit clone https://github.com/ggml-org/ggml
Claude Codeskills systembughunter83 skillspatterns681 reportscommands15 slashBurp MCPengagements
+98 that day★ 6.5KPython

Claude Code skill bundle for bug hunting and red-team work

claude-bughunter is a self-contained skill bundle for Claude Code, for bug hunting and external red-team work. The README lists 83 skills and 15 slash commands.

  • Holds 681 patterns from disclosed reports across 24 core vulnerability classes; 433 of them are individually cited and auditable.
  • Includes enterprise identity and infrastructure attack matrices, engagement-folder scaffolding, and Burp MCP integration.
  • The README says it was used in authorized red-team and bug-hunting engagements and on public training platforms such as DVWA and OWASP Juice Shop.
coding agentold web habitsweb skillsbest practicesmodern webplatform APIscompat databrowsers
+1.2K that day★ 7.3KTypeScript

Skills that steer coding agents to modern web practices

The source repo for GoogleChrome/modern-web-guidance, with code and docs for authoring, calibrating, and evaluating its content. It holds skills that bring web platform expertise, best practices, and browser compatibility data to coding agents.

  • Steers agents away from legacy patterns toward modern web platform capabilities.
  • Stewarded by the Google Chrome, Microsoft Edge, Mozilla Firefox, and MDN teams with the web community.
  • Includes an eval dashboard plus contributing and governance docs.
trynpx modern-web-guidance@latest install
market datacalendartrade skillsClaude Skillsreview + planrisk, journalinvestordecides
+706 that day★ 5.6KPython

Claude skills for investors: market review and risk management

Claude Trading Skills is a Claude Skills-based trading workflow toolkit for time-constrained individual investors. It aims to structure market review, risk management, trade planning, journaling, and continuous improvement.

  • Designed for long-term investors with ETFs and dividend stocks as the core, plus disciplined swing trading as a satellite.
  • The description lists market analysis, technical charting, economic calendars, screeners, and strategy development.
  • The README states it is not a signal service or a promise of profitability; buy and sell decisions are not outsourced to AI.
agent skillfolderevaluatorskillevaluatorTier 1 checksno API keywith/withoutskill runs
+122 that day★ 1.8KPython

Evaluate AI agent skills and their effect on task results

SkillEvaluator checks AI agent skills and measures their effect on task results: start with quality checks, then compare runs with and without a skill. Part of the NVIDIA Verified Skills pipeline.

  • One CLI, three ways in: run it directly, add it to your repo via CI, or wire it to your agent.
  • Tier 1 checks run without an API key; the repo description also mentions semantic overlap detection and synthetic evaluation datasets.
  • skillevaluator ci init opens a setup PR on GitHub.
trypip install skillevaluator