GitHub 2026-05-31 Weekly Trending Projects

An engineering‑oriented tool for short‑video production that integrates multiple LLMs and voice synthesis, offering Web UI, API, batch generation and subtitle/music management. Suited for teams automating content output, though deployment complexity and compliance require additional effort.

Understand-Anything is a plugin-based tool that converts codebases and knowledge bases into interactive knowledge graphs using multi-agent LLM extraction; it offers a visual dashboard, semantic search and diff impact analysis—suited for engineering and product teams needing fast onboarding and systemic understanding.

A hands-on, end-to-end AI engineering curriculum for programmers and engineers that spans math to multi-agent production; it stresses building algorithms from first principles and ships reusable artifacts each lesson to bridge theory and production.

Role-focused customizable plugins to integrate tools and codify workflows for Claude, improving team consistency.

CodeGraph provides a local semantic index and graph traversal layer for Claude Code, dramatically reducing file scanning and tool calls—ideal for privacy-minded teams and large/multi-language codebases to accelerate code exploration and impact analysis.

Stop Slop is a compact LLM skill and reference set that identifies and removes AI-generated stylistic tells—phrases, structural clichés, and sentence-level patterns—helping prompt engineers, editors, and tooling teams make prose more natural and editor-ready in review and automation workflows.

Organizes 754 production-grade security skills in agentskills.io format mapped to five major frameworks, enabling rapid conversion of AI agents into actionable security analysts; license and maintenance considerations should be resolved prior to enterprise adoption.

Taste Skill supplies composable frontend 'skills' and an image-to-code pipeline that use adjustable design dials and GSAP scaffolds to accelerate turning AI outputs into implementations; however, missing license and contribution metadata require compliance and maintenance validation before production adoption.

MarkItDown converts varied documents into structured Markdown for LLM pipelines, enabling automated text workflows while requiring care with optional dependencies and plugin compatibility.

Anthropic-maintained Claude Code plugin directory offering standardized plugin structure and easy installation; useful for developers and operators, but third-party plugins require careful security and license review.

ECC is a production-focused AI agent operator system offering cross-harness skill libraries, memory optimization, continuous learning, and security tooling—designed for engineering and research teams deploying agent orchestration at scale.

A Cursor plugins marketplace centered on standardized plugin manifests and developer-tooling patterns; useful for building internal tools and automation references, but verify activity and compatibility before production use.

Agent governance: policy-as-code enforcement, auditable logs, and least-privilege controls for production agent deployments.

Dograh is an open, self-hostable voice-AI platform for engineering teams, offering a drag-and-drop workflow builder, pluggable LLM/STT/TTS, and telephony integrations — suited for production voice agents where data residency and source-code control matter.

Harness is a team-architecture factory for Claude Code that converts domain descriptions into six types of agent teams and matching skills, enabling rapid construction, orchestration, and validation of multi-role automated workflows.

Herdr delivers agent-aware multiplexing and persistent sessions inside your existing terminal, showing real process views and supporting detach/reattach and remote attach—suited for developers and operators who need to observe and manage long-running background agents.

An IDE-grade local coding agent that provides LSP/DAP coupling, a rich toolchain and subagent collaboration—well suited for large codebases needing precise edits, debugging, and automated code review.

Heretic pairs directional ablation with TPE-based automatic tuning to offer a one‑click, low‑damage decensorship tool for GPU-equipped researchers and engineers—useful for probing model semantics and generative boundaries, but requiring careful legal and misuse risk assessment.

Harness provides teams using Claude Code with an auditable delivery loop that ties planning, implementation, verification, review and release into a controlled workflow—suited for engineering teams that require governed agent development and reproducible release evidence.