GitHub 2026-02-09 Weekly Trending Projects

Shannon is a white‑box autonomous AI pentester that, from source‑aware analysis, constructs and executes real exploit PoCs and produces reproducible reports—enabling security and development teams to move pentesting from a yearly event to a continuous process.

OpenClaw is a local-first personal AI assistant that runs on your devices, offering multi-channel access, voice wake and a live Canvas — suited for privacy-focused users and developers seeking deep integrations.

Provides a discoverable, installable catalog of agent skills for Codex to help teams package and reuse task capabilities; the repository currently lacks license information and active maintenance, so it is best evaluated and customized within controlled environments.

For developers and teams using Claude Code: automatically capture, semantically compress, and persist session outputs, then inject relevant context into future sessions to improve continuity and retrieval efficiency.

Pi is a developer-focused monorepo toolkit offering a unified multi-provider LLM API, agent runtime, coding agent CLI, terminal and web UI components, and tools to manage vLLM GPU pods—suited for teams with TypeScript/Node skills building and deploying custom AI agents.

Provides 13 Claude Code hook examples, sub-agents and team validation to enable controllable, auditable AI agent workflows.

QMD is a privacy-oriented on-device document retrieval tool that combines BM25 full-text, vector semantic search and LLM reranking — suited for individuals or teams needing offline, controllable search with structured outputs for agent workflows.

An experimental AI agent for experienced Neovim users that provides controlled code assistance via customizable SKILL/rules and cmp completion—suitable for developers willing to accept alpha-stage risk for trial and extension.

ChatDev 2.0 is a zero-code multi-agent orchestration platform combining research-grade methods (e.g. MacNet, puppeteer) to let users configure agents and workflows to rapidly build, run and scale complex automation — while requiring attention to licensing, runtime costs and security constraints.

Maestro is a power‑user cross‑platform desktop orchestrator combining parallel multi‑agent workflows, Git worktrees isolation, Auto Run playbooks and mobile remote control, designed for managing long‑running, high‑concurrency AI automation with visualization.

nanochat is a minimal, hackable end‑to‑end LLM training and chat framework that can rapidly reproduce GPT‑2‑grade models on a single machine or an 8×H100 node; it is well suited for research, teaching, and fast experimentation, but pay attention to license, contributor metadata, and hardware costs.

PageIndex converts long documents into semantic trees and uses LLM-driven tree search for reasoning-based retrieval, delivering vector-free, high-precision, and explainable retrieval for professional domains like finance and law.

Playwright CLI delivers a token‑efficient Playwright command‑line interface for coding agents to run browser automation and testing within constrained LLM contexts, suited for high‑throughput automation.

Prek reimplements pre-commit in Rust, offering a single dependency-free binary, shared toolchains, and parallelized installs/execution to speed setup and hook runs; suitable for teams seeking faster, disk-efficient pre-commit alternatives, but pay attention to partial language parity and the repository's current maintenance indicators.