01Open source
Repos
Open-source repos worth knowing, by the problem they take on.
AI & Agents
AI Anti-Laziness Skill
A skill for AI agents that enforces completion discipline using a Depth Tree method, splitting tasks into N layers and giving each leaf the full time budget to counter laziness and premature completion. It is notable for its research-grounded approach and runnable gates for Claude Code.
Unified Harness Router
Self-hosted Apache-2.0 gateway that exposes Codex, Claude Code, Hermes, PI, DSH, and other agent harnesses through one API with sessions, streaming, files, cancellation, and failure handling. It implements the open Unified Harness Protocol to address harness fragmentation.
Autoprompt Coding Skill
A coding-agent skill that claims to reduce failures by 45% on agentic coding tasks. It installs via npm or GitHub Releases and provides benchmarks, run controls, and workflow guidance for coding agents.
De-AI Writing Skill
A portable De-AI writing skill for Agent Skills-compatible agents, with plugins for Claude Code, Codex, Grok Build, and Antigravity. It repairs narrative architecture in fiction and applies venue-matched rules to professional prose, based on StoryScope research.
AI Scope Guard
A multi-platform hook and skill guard for AI coding agents in Codex, Claude Code, OpenCode, Hermes, and Pi workflows that blocks unrequested hashes, checksums, and scope creep. It targets overengineering behaviors like unnecessary hardening and task expansion.
Offline Durable Memory
An offline durable memory tool for Claude Code and Opencode that logs sessions locally and condenses them into a resume-ready summary using a classical Python summarizer, with no API key or external model. It reduces token waste from re-explaining projects each session.
Agent Long-Term Memory
A local long-term memory runtime for AI agents that uses plain Markdown as the source of truth and ranked local retrieval, with a sleep-time management layer. Claude Code and Codex share one store without an API key.
Dev Tools & CLI
Remote Command Monitor
A local-first command monitor that lets you prefix commands with hao and follow their live output and finish notifications from your phone. Useful for server admins and anyone running long tasks like model training.
Learn PyTorch
An educational project for learning PyTorch by building your own implementation, inspired by Harvard's TinyTorch and structured as 20 progressive modules. It is a learning resource rather than a production tool.
Local Ops Console
A macOS local web console for starting and monitoring local services and batch tasks, with a zero-dependency Python backend and no-build frontend bound to loopback. It is single-user and currently in alpha, not for remote or multi-user use.
Linux App Uninstaller
A universal Linux application uninstaller and deep-clean residual purger that removes orphaned user-space directories like ~/.config, ~/.cache, ~/.local/share, and Flatpak app folders after package removal. It is a TUI utility for cleaning leftover files.
Postgres Terminal
An interactive terminal UI for monitoring multiple Postgres databases in one place, described as htop for Postgres and powered by pgbot's read-only diagnostics. Aimed at database developers who want a modern, focused overview.
Finance & Quant
IB Gateway Docker
ib-gateway-docker packages Interactive Brokers Gateway with IBC and Xvfb so it runs unattended in a container. It is widely useful for automated trading infrastructure that needs a stable IB connection.
CZSC Technical Analysis
czsc is a Chan theory technical analysis toolkit for Chinese stocks and futures, with core algorithms (fractals, strokes, centers) implemented in Rust and exposed to Python via PyO3. It is an established, actively maintained niche tool for quant trading.
Financial Charting Library
Vela is a headless interactive financial charting library with a native WebGL2/Canvas2D renderer and optional workspace UI. It is notable for its plugin SDK and active development by established provider LuxAlgo.
Quantitative Finance Framework
FinHack is an extensible quantitative finance framework covering data collection, factor calculation and mining, machine learning, strategy writing, backtesting, and live trading. It supports A-shares, US stocks, futures, and crypto, with DuckDB and minute-level backtesting planned.
Mobile
Mobile Terminal Workspace
Conduit is a mobile terminal workspace for Android and iOS that provides SSH, Mosh, SFTP, hardware-key authentication, tmux helpers, and an optional local shell. It keeps hosts, keys, and fingerprints on-device with no account or cloud sync.
Video Search SDK
V-Modal SDK for Android is a Kotlin library adding multimodal video/image search and mobile-friendly uploads to video apps. It provides a small typed API; the app owns screens and lifecycle while the SDK handles gateway, parsing, signed uploads, and progress.
iPhone Mirror
iPhoneMirror is a Windows tool for low-latency iPhone and iPad screen and system audio mirroring over USB or AirPlay, with D3D11 preview, multi-device sessions, and OBS output. It is a public preview and currently supports Windows x64 only.
Rust Android VPN
MSN-GUARD is a native Android VPN client with a Rust core that tunnels all device traffic through five protocols: SHARD, MASQUE/HTTP-3, WireGuard, WARP-on-WARP, Psiphon, and Tor. It operates at Android's VpnService level, covering TCP, UDP, and QUIC from any app.
iPhone Location Control
Roam Control is a source-available SwiftUI app for choosing, testing, and moving an iPhone's reported location from an Apple Maps interface. It supports fixed locations, walking routes, favorites, history, and on-device pairing for development and QA testing.
Compose Navigation Graph
Compose Navigation Graph is an Android Studio plugin that visualizes Compose navigation as an interactive map of rendered previews, arguments, and transitions. It statically extracts the graph across modules and supports pull-request validation via a committed .nav baseline.
Robotics
CoC Autolabeler
NVIDIA CoC Autolabeler is an open pipeline that generates Chain-of-Causation reasoning labels and meta-actions from raw autonomous driving clips. It is notable for producing causal labels for autonomous vehicle data.
Show-Harness VLM Agent
Show-Harness is a harness for evaluating vision-language model agents on robot control tasks, with models and a dataset. It is notable as a fresh evaluation framework from Show Lab with high momentum.
VLAct VLA Pre-training
VLAct is the official code for representation-centric continued pre-training of vision-language-action models, with a continued-pretraining backbone and downstream checkpoints. It is notable for focusing on representation quality rather than data scaling for robot learning.
POLI Point Cloud
POLI is a self-supervised network that predicts a Gaussian ellipsoid per LiDAR point to capture local geometry on a statistical manifold. It is notable as RSS 2026 research code for improving 3D perception without labels or architecture changes.
OpenWAM World Model
OpenWAM is an open research stack for modular World-Action Model development, including infrastructure for composing model, representation, training, and evaluation choices, controlled studies, and a pretrained model. Notable for systematic world-action pretraining for generalist robot policies.
verl-vla Post-training Framework
verl-vla is a unified cloud-edge post-training framework for vision-language-action policies, built on verl, connecting distributed resources for human-in-the-loop data collection, fine-tuning, evaluation, and reinforcement learning. It is notable for providing a continuous post-training path across these stages.
WNM-3D Navigation Model
WNM-3D is the official implementation of a world navigation model with 3D scene conditioning for closed-loop vision-and-language navigation. It is notable as a 3D-conditioned navigation model for VLN tasks.
Science & Simulation
OpenQuake Engine
OpenQuake Engine is open-source software for calculating and assessing seismic hazard, risk, and decision-making tools developed by the GEM Foundation. It is widely used, actively maintained, and substantial.
PySPH Framework
PySPH is a Python framework for Smoothed Particle Hydrodynamics simulations, with performance-critical parts in Cython and OpenCL. It is a real simulation engine that supports OpenMP, OpenCL, and MPI and is actively maintained.
FiPy PDE Solver
FiPy is a finite volume partial differential equation solver written in Python. It is a NIST-maintained research code that remains actively developed for solving PDEs.
Systems & Languages
Rspack Bundler
Rspack is a fast Rust-based web bundler that modernizes the webpack API for compatibility with existing plugins and loaders. It offers significantly faster builds and hot module replacement than webpack.
C3 Compiler
c3c is the compiler for C3, a general-purpose systems programming language that modernizes C with familiar syntax and semantics. It targets fast native software and is actively developed.
Lightning CSS
Lightning CSS is an extremely fast CSS parser, transformer, bundler, and minifier written in Rust. It provides typed property values and can be used standalone, with Parcel, or via plugins for other tools.
Scala 3 Compiler
Scala 3, also known as Dotty, is the compiler, standard library, and language specification for the Scala 3 programming language. It is a major, actively developed language compiler.
Wazero WASM Runtime
wazero is a zero-dependency WebAssembly runtime written in Go, compliant with Core Specification 1.0 and 2.0, and does not rely on CGO. It lets Go developers embed and run Wasm modules while preserving cross-compilation.
Carbon Language
Carbon Language's main repository contains design documents, implementation, and tools for an experimental C++ successor from Google. It is not ready for production use but is actively developed with high potential.
Babel Compiler
Babel is a compiler that transforms next-generation JavaScript into backward-compatible code for current environments. It is a widely used, community-driven core tool in web development.
This Week's Movers
Android Automation Tool
ARTEMIS converts natural-language instructions into Android automation, executing cross-app workflows and capturing logs via MCP for coding assistants. It reports 99%+ success on the AndroidWorld benchmark.