Forkcast Today #2: openhuman, agentmemory, superpowers, AiToEarn, telegraf, react-doctor

Episode Summary#

This episode covers six trending projects spanning personal AI assistants, agent infrastructure, content automation, and code quality. Highlights include OpenHuman’s local-first AI assistant with 118+ integrations, the Superpowers agentic skills framework that has taken GitHub by storm, an AI content marketing platform for creators, and a React code quality tool that catches bad patterns before they ship.

Repositories Covered#

openhuman#

A local-first personal AI assistant with memory tree, Obsidian wiki integration, 118+ third-party integrations, and TokenJuice compression. Runs with Rust, React, and Tauri for a native desktop experience.

agentmemory#

A persistent memory system for AI coding agents that maintains context across sessions. Works with Claude Code, Codex, Cursor, and Copilot without client-specific dependencies.

superpowers#

An agentic skills framework and software development methodology that enforces TDD, spec-first workflows, git worktree isolation, and two-stage code review. With 200K+ stars, it defines how AI coding agents should work reliably across Claude Code, Codex, Cursor, and more.

AiToEarn#

An open-source AI content marketing agent platform that automates the entire content pipeline — from creation to publishing across 14+ platforms (TikTok, YouTube, Xiaohongshu, Bilibili, and more) to engagement and monetization. Built with Next.js, Node.js, and Electron.

telegraf#

InfluxData’s plugin-driven server agent for collecting, processing, aggregating, and writing metrics and logs. Part of the TICK stack with 300+ plugins covering systems, cloud services, databases, and message queues.

react-doctor#

A code quality tool from Million.js that detects poor React patterns and anti-patterns in your codebase. Catches issues like unnecessary re-renders, missing memoization, and inefficient component structures before they impact production performance.

scientific-agent-skills#

A collection of specialized AI agent skills for scientific research workflows. Covers literature review, citation graph analysis, peer review simulation, and methodology validation to accelerate academic and research processes.

Personal_AI_Infrastructure#

Your AI tools know nothing about your life priorities, your projects, or what matters to you. Every conversation starts from zero. Positioned as a Life Operating System. Pulse daemon on port 31337 with 22 API routes, voice, hooks, observability, cron, wiki API. Algorithm v6.3.0: 7-phase OBSERVE→THINK→PLAN→BUILD→EXECUTE→VERIFY→LEARN loop. 45 public skills, 171 workflows, 37 hooks. ISA (Ideal State Artifact) with 12 sections and 5 identities. Memory v7.6: WORK, KNOWLEDGE, LEARNING, RELATIONSHIP, OBSERVABILITY, STATE. No RAG by design — uses filesystem + ripgrep. 12 security gates on every release. MIT licensed.

supertonic#

Every TTS service is cloud-locked, expensive at scale, and fails on edge cases like financial expressions or phone numbers. You need TTS that runs offline and gets the details right. 31 languages in v3. ~99M parameters — much smaller than 0.7B-2B class open TTS systems. ONNX Runtime for cross-platform inference. Raspberry Pi demo at 0.3× RTF with zero network. Browser via onnxruntime-web (WebGPU/WASM). Handles financial expressions, phone numbers, technical units correctly where ElevenLabs, OpenAI, Gemini, and Microsoft TTS all fail. 11 inference runtimes: Python, Node.js, Browser, Java, C++, C#, Go, Swift, iOS, Rust, Flutter. Expressive tags: laugh, breath, sigh. MIT + OpenRAIL-M licensed.

CloakBrowser#

Every web scraper eventually hits Cloudflare, reCAPTCHA, or fingerprinting. Stock Playwright scores 0.1 on reCAPTCHA v3 — instantly flagged as a bot. 57 C++ source-level fingerprint patches (not JS injection). reCAPTCHA v3 score: 0.9 (human-level) vs Playwright’s 0.1. Cloudflare Turnstile: PASS. FingerprintJS: PASS. BrowserScan: NORMAL (4/4 checks). bot.incolumitas.com: 1 fail vs Playwright’s 13 fails. 30+ detection sites tested. navigator.webdriver: false (source-level patch). Humanize mode: Bezier mouse curves, per-character typing delays, realistic scroll acceleration. ~190MB RAM idle, ~280MB with 3 tabs. Drop-in Playwright replacement. Cross-platform: Linux, macOS, Windows. MIT wrapper + proprietary binary.

GTweak#

Windows ships with telemetry, bloatware, and dozens of background services that eat CPU cycles and compromise privacy. Disabling them manually means digging through registry and group policy. Windows 10+ support, .NET Framework 4.8. 98 releases, latest v5.4.9 (April 2026). Activation via HWID and KMS. Disables Defender, SmartScreen, VBS, UAC. Blocks Microsoft shadow domains via hosts and firewall. Removes OneDrive, Edge, Cortana, Copilot, Recall. Applies ultimate performance power plan. Audio driver power settings fix for Realtek. NTFS compression, RAM/temp clearing, secure deletion. Executes custom .ps1/.cmd/.bat/.reg scripts with TrustedInstaller privileges. 12+ language translations. BSD-3-Clause.

skills#

A collection of production-grade engineering skills for AI coding agents by Matt Pocock. Provides structured, reusable capability files that make agents more reliable at TypeScript, testing, and general software engineering tasks.

brush#

3D reconstruction tools are locked behind expensive proprietary software. Researchers and indie devs have no fast, open path from photos to 3D models. Gaussian splatting engine in Rust. Cross-platform: macOS, Windows, Linux, Android, and browser via WebGPU. Training and rendering faster than gsplat (per author benchmarks). Real-time interaction during training with live view updates. Supports COLMAP data, Nerfstudio-format datasets, .ply and .compressed.ply files. 4D scenes with delta frames compatible with cat-4D and Cap4D. Burn ML framework for GPU-agnostic compute. Build: Rust 1.88+. WASM via wasm-pack, Chrome 134+ only. 1,154 commits. Apache-2.0 licensed.

How-To-Secure-A-Linux-Server#

Most Linux server hardening guides are either outdated within months or written by vendors pushing their own security products. You need an evolving, community-vetted reference. Dedicated files for kernel sysctl hardening and nginx security. 12 topic tags covering server hardening, security hardening, Linux server configuration. 269 commits of community-reviewed updates. CC-BY-SA-4.0 licensed — freely shareable and remixable. 1.8k forks from the community. No dependencies, no tools to install — pure knowledge. Continuously evolving with community contributions.

hysteria#

Standard VPN protocols get throttled or blocked on restricted networks. TCP-based proxies choke on lossy connections with 5%+ packet loss, making them unusable in many regions. Customized QUIC protocol designed for unreliable and lossy networks. Traffic masquerades as standard HTTP/3 to evade censorship without collateral damage. Proxy modes: SOCKS5, HTTP Proxy, TCP/UDP Forwarding, Linux TProxy, TUN. Builds for every major platform and architecture. v2.9.1 released May 10, 2026. 81 releases, 1,310 commits. MIT licensed. Legacy v1 maintained separately. Custom authentication, traffic statistics, access control.

LLMs-from-scratch#

LLMs feel like magic black boxes. Without understanding the internals of attention, pretraining, and finetuning, you’re just guessing at prompts. 7 chapters + 5 appendices. Implements GPT, Llama 3.2, Qwen3 (dense + MoE), Gemma 3, Gemma 4, Olmo 3, Tiny Aya from scratch. BPE tokenizer built from scratch. Covers KV Cache, GQA, MLA, Sliding Window Attention, Gated DeltaNet, MoE, FLOPs analysis, DPO alignment, LoRA (Appendix E), DDP (Appendix A). Companion 170-page PDF with 30 quiz questions per chapter. 17-hour 15-minute video course. Code runs on conventional laptops with auto GPU detection. 94.2k stars, 14.5k forks. Published 2024, ISBN 978-1633437166.

acton#

TON smart contract development requires juggling 5+ separate tools: one for scaffolding, one for building, one for testing, one for deployment, and one for verification. All-in-one CLI in Rust: scaffolding, build, testing, scripting, wallet/network operations, verification, linting, formatting, debugging, low-level VM tooling. Tolk-first workflow with built-in wrappers. Auto-generated TypeScript wrappers. Test runner with fork mode, gas snapshots, coverage, mutation testing, fuzzing, browser test UI. macOS (ARM64/x86_64) and Linux (x86_64/ARM64) first-class support. Windows via WSL. 38 releases, v1.0.0 released May 11, 2026. 2,215 commits. Dual Apache-2.0/MIT licensed.

cua#

Computer-Use agents need sandboxes to operate safely, but setting up isolated VMs for every agent session is slow and brittle. You need infrastructure that just works across OSes. Cua Driver: drives native macOS apps in background without stealing cursor or focus. Works on non-AX surfaces: Chromium, Blender, Figma, DAWs, game engines. Cua Sandbox: 5 OS targets (Linux container, Linux VM, macOS, Windows, Android + BYOI). Cua-Bench: evaluates on OSWorld, ScreenSpot, Windows Arena. 6 packages: cuabot, cua-agent, cua-sandbox, cua-computer-server, cua-bench, lume. Lume: macOS/Linux VM management on Apple Silicon via Virtualization.Framework. 482 releases, 3,236 commits. MIT licensed.

spec-kit#

GitHub’s official toolkit for spec-driven development that generates AI-assisted PRDs before any code is written. Integrates with Copilot to turn specs into implementation plans.

Watch#

• Video: https://www.youtube.com/watch?v=N98OKwdXiwY
• Cover: https://i.ytimg.com/vi/N98OKwdXiwY/hqdefault.jpg

Notes#

Transcript and notes will be added from Forkcast output artifacts.