GitHub 2026-07-21 Weekly Trending Projects

Hallmark is an LLM design skill for Claude, Cursor, and Codex that generates self-contained HTML/CSS pages using macrostructures, 20 themes, and 57 quality gates—suited for teams needing portable design templates and controlled AI output; however, repository activity and compatibility should be evaluated before production use.

An open-source cross-platform video editor rewrite targeting creators and developers: Rust core with a plugin-first design to enable desktop/browser/mobile parity, scripting, and headless automation. Currently in active rewrite and unstable—suited for technical early adopters and integration experiments.

Aggregates numerous LLM application examples and links with emphasis on Agents and RAG practices; well suited for learning, comparison and rapid prototyping, but users should watch entry quality and maintenance stability.

DeepTutor centers on an agent-native architecture that combines TutorBots, persistent memory, and RAG knowledge bases to deliver an extensible personalized tutoring and collaborative workspace—well suited for autonomous learning, teaching automation, and knowledge-driven conversational education products.

dcg provides real-time destructive-command protection for AI coding agents, blocking with low latency and offering safer alternatives; suited for local development and CI pipelines.

Vibe-Trading is an agent-centric trading platform for quant researchers and developers, offering connector-first broker access, paper-account testing across multiple brokers and a mandate/audit safety model for constrained live orders. It is suited for rapid LLM-driven strategy experimentation, but live use requires careful compliance and risk review.

code-review-graph builds an incremental code-structure graph with Tree-sitter to compute change blast-radii and minimize the context AI reviewers must read; suited for large repositories seeking to reduce token/read costs and integrate with CI/IDE.

UI Skills provides an npx-driven CLI to route and retrieve categorized UI skill sets for design engineers, enabling quick discovery and reuse; repository maintenance and metadata should be verified before adoption.

An extensible collection of agent skills for product and engineering teams that automates PRDs, planning, issue breakdowns and TDD flows, enabling modular integration of agent capabilities into repositories.

Transforms long-form content into callable, testable AI skills with indexes and tests for agent reuse.

OfficeCLI delivers a single-binary, cross-platform solution for AI agents and developers to read, write and render Office documents with high fidelity, suited for automated generation, review and programmatic editing workflows.

Codex CLI is a Rust-centered terminal coding agent providing cross-platform binaries and ChatGPT integration, suited for developers and teams who want command-line AI coding assistance locally.

Kimi Code CLI is a terminal-first AI agent for developers that combines ACP/MCP and Zsh support for in-terminal code editing, shell execution and automation; evaluate project maturity and licensing before production adoption.

Open Interpreter is a Rust-based Codex derivative coding agent that uses switchable harnesses and local sandboxing to optimize low-cost models for real-world code generation and automated testing, suitable for developers and teams wanting local agent deployment.

A tool that converts natural-language descriptions into self-contained, theme-aware, high-resolution architecture and flow diagrams—well suited for rapid documentation and communication, but limited by an unclear license and minimal community activity.

pi is an extensible TypeScript coding-agent platform with multi-LLM support, tool-calling and a CLI for automated coding workflows.

A curated hands-on tutorial collection to learn systems by reimplementing technologies from scratch.

This repository collects Anthropic's 'Code with Claude' hands-on workshops covering model selection, multi-agent design, memory and evaluation—useful for learning agent engineering and eval-driven development—but it is declared unmaintained and heavily relies on Anthropic services, posing migration and long-term use risks.