Agent Skills is an open-source collection of engineering workflows for AI coding agents, developed by Addy Osmani. It packages senior-engineering practices into skills covering the development lifecycle: defining specifications, planning, incremental implementation, testing and debugging, constraints, code review, web-performance auditing, code simplification, and shipping. Slash commands such as /spec, /plan, /build, /test, /review, and /ship activate the corresponding workflows, while skills can also activate automatically based on the work being performed. The /build workflow can generate a plan and implement its tasks in an approved pass, while retaining test-driven verification, individual commits, and pauses for failures or risky steps. It can be installed with the skills CLI or integrated into supported coding agents including Claude Code, Cursor, Codex, Copilot, and Cline.
AgentVerse OS is a personal cloud operating system for a developer and their AI agents, installed on an Ubuntu server and accessed through a browser from a laptop, tablet, or phone. It provides a windowed desktop, isolated project workspaces, browser-based VS Code and a terminal, and supports Claude Code and Codex agents whose work remains on the server. Each project runs in an isolated Incus container with Docker inside, its own network and access gate. A Rust core coordinates projects, workspaces, capabilities, applications, backups, and updates; Coder controls the workspaces, Komodo deploys application stacks, Caddy provides the entry points, and Tailscale supplies private access and certificates. Projects request capabilities such as S3 storage, an LLM gateway, or notifications, which the core connects to provider networks and exposes through environment variables. The system includes a store combining catalogs from Runtipi, Coolify, and Umbrel, with guided installation, logs, snapshots, rollback, and application updates. It also supports ZFS snapshots, restic backups, an embedded Svelte desktop PWA, and one-command installation from a release package or source. The project is in alpha, is designed for one person, has no user accounts or permissions yet, and requires a clean Ubuntu installation and a Tailscale account. The repository is licensed under Apache-2.0; store applications retain their own licenses.
Anki is a spaced repetition flashcard program. This repository contains the source code for its computer version, developed by the Anki project.
AuK is an open-source 1.5-billion-parameter foundation model for speech generation and editing, developed by Tencent Hunyuan. It exposes zero-shot and instruction-based text-to-speech, content and lyric rewriting, pitch, speed, volume, emotion, timbre, accent, whisper and nonverbal-sound editing, speech enhancement, speaker separation, music separation, and target-speaker extraction through a unified natural-language instruction interface. Inputs can include text instructions and optional source or reference audio; outputs are generated or edited audio. The repository provides an AuK base model for configurable high-quality generation and AuK-Flash, a distilled variant using four-step inference, along with command-line, Python, Gradio, ComfyUI, and fine-tuning interfaces. The code is released under the MIT License, and model weights are distributed through Hugging Face and ModelScope.
Bend 2 is a parallel programming language created by Victor Taelin and contributors, designed for AI-written code that must preserve declared rules. Its Python-like syntax includes dependent types, purity, linearity, and explicit proofs. A project can define invariants in a `LAWS.bend` file and provide proofs in `PROOF.bend`; the compiler checks those proofs when code changes, blocking implementations that violate the declared laws. Bend uses divide-and-conquer parallelism rather than manually managed threads, locks, or GPU kernels, and can target CPUs and GPUs through C, CUDA, and Metal, with JavaScript also supported. It is distributed through an install script and is positioned primarily for Linux and macOS; the repository describes it as a young language with limitations including no Windows support, no proof search, slow native compilation, and experimental areas of the compiler and runtime.
Birdview is an open-source developer tool that makes AI coding agents map a project's architecture before editing. Its Birdview workflow defines modules, responsibilities, file ownership, evidence, relationships, and layout in an architecture JSON file; an agent then declares planned task scope, targets, files, lifecycle phases, and verification records against that map in an activity JSONL stream. Validators check the schemas and cross-record rules before a renderer produces a standalone interactive HTML view with architecture, changes, comparison, module evidence, and activity-history views. Birdview supports automatic or on-demand activation through project-level agent guidance and includes Node.js scripts, JSON Schemas, example maps, and tests. The generated output has no server or network dependency, but the v0.1 implementation is file-based: activity is agent-declared rather than automatically observed, and updates require regenerating the HTML. It is MIT-licensed and is not published to npm.
Claude Code is Anthropic's agentic coding tool for the terminal, IDEs, and GitHub. It uses natural-language commands to understand a codebase, create and read files, execute commands, run tests, explain code, manage Git workflows, and handle routine development tasks. It can also load persistent project context, run custom slash commands, use plugins with custom commands and agents, and operate with configurable autonomy while leaving actions such as final pull-request merging to a human. The official repository documents installation for macOS, Linux, and Windows, and identifies npm installation as deprecated.
An open-source Codex configuration that uses Astra as the root orchestrator and independent reviewer, with Luna-based subagents for exploration, implementation, testing, and research. It provides separate Pro and Plus profiles, TOML role configurations, an Astra orchestrator skill, project-level AGENTS.md instructions, and shell and PowerShell installers that copy the selected configuration into a target repository. The profiles set model, reasoning-effort, approval, sandbox, and concurrent-thread settings, while named role files can override the defaults. The repository also includes guides for orchestration patterns, plan-specific setup, iteration, and token-usage reporting from Codex session logs. It is licensed under Apache License 2.0.
GitDiagram is a web application that turns public or private GitHub repositories into interactive architecture diagrams with links back to the source files and directories. It retrieves a repository's default branch, tree, README, and bounded, integrity-checked source excerpts through the GitHub API; an OpenAI or OpenRouter model produces a source-grounded overview and graph, whose groups, nodes, edges, shapes, labels, and paths are validated against the repository. A deterministic compiler converts the validated graph to Mermaid, while the browser sanitizes and renders the result and can export Mermaid source or a PNG. Successful diagrams are persisted for later reopening, and private-repository access uses a GitHub token supplied for the relevant request rather than embedding the token in public links.
OpenSpec is an open-source, tool-agnostic spec-driven development framework from Fission AI for AI coding assistants. It stores requirements and development artifacts as plain Markdown files in a repository, using an `openspec/` structure containing specifications and proposed changes. Its workflow supports exploratory discussions, `/opsx:propose` for generating a proposal, requirements and scenarios, a technical design, and implementation tasks, followed by `/opsx:apply` to implement the tasks and `/opsx:archive` to update the specifications. Coding agents can interact with it through slash commands or an MCP server. The project is designed for iterative, brownfield development and supports shared or cross-repository requirements through beta Stores. The CLI requires Node.js 20.19.0 or higher and is installed with npm.
page-mascot is a web component that adds an interactive character which tracks the cursor and displays a reaction when clicked. Each character uses two aligned 3×3 sprite sheets: one with nine head directions and one with nine expressions. The pointer angle selects a direction, with a dead zone near the cursor, while a click displays a reaction for half a second. Existing mascots can be supplied as image sheets, or the included coding-agent skill can generate nine directions and nine expressions for a custom character and place it on a page. It supports colour, ink, sketch, riso, paper, and pixel styles; tracking is disabled without a fine pointer, and click animation respects prefers-reduced-motion. The package is distributed through npm and licensed under the MIT License.
Rune is a GPU-rendered, keyboard-driven IDE for power users, developed by Unstable Build. It combines code editing, terminals, command-line tools, language intelligence, debugging, file and workspace management, and AI agents in a composable multi-workspace environment with a character-grid interface. Rune Agent is provided separately as an extension in the same repository rather than being part of the core editor. Rune is an application written in Go and built with standard Go tooling; its GPU renderer uses cgo and requires platform-specific graphics and development libraries. The application is licensed under the GNU General Public License version 3 or later, while the separate rune-go-sdk extension module is licensed under Apache-2.0.
RustFS is an open-source, distributed object storage system built in Rust. It provides S3-compatible storage for data lakes, AI, and big-data workloads, with single-node and distributed deployments, versioning, object lock, server-side encryption, replication, lifecycle management, quotas, event notifications, IAM policies, audit logging, and a web console. It also supports the OpenStack Swift API with Keystone authentication, plus SFTP, FTPS, and WebDAV interfaces. RustFS can be deployed from binaries, Docker or Podman, Docker Compose, Kubernetes through Helm, Nix, or from source. Its S3 compatibility is tracked in a compatibility matrix, and selected features such as S3 Tables through an Iceberg REST catalog and MinIO on-disk compatibility are marked as preview. The project is licensed under Apache 2.0.
security-audit is a Cloudflare coding-agent skill that orchestrates multi-phase security audits of codebases. It uses reconnaissance to map architecture, trust boundaries, input surfaces, prior evidence, and deterministic coverage; assigns isolated agents to coverage-led hunting; sends each unique candidate to a fresh verifier; and records confirmed, needs_validation, and rejected results in machine-readable findings.json validated against a report schema. Independent agents verify final source claims, after which the skill derives target-neutral reports and coverage records. It can be installed with the Skills CLI and requires a tool-using coding agent with parallel sub-agent support, Node.js for its validators, and an OS-enforced sandbox for target-controlled execution. The repository is licensed under MIT.
SoL-Pi is an open-source standalone extension for the Pi coding agent, developed and maintained by NVIDIA. It packages four opt-in efficiency mechanisms: Action Fusion runs a follow-up validation command in the same edit or write call; ObservationPack replaces repeated large tool results with stable handles that support exact paged recall; the Evidence-Preserving Reducer condenses diagnostic logs into receipts only when retained quotations match archived source material; and Online Context Compact marks completed plan steps for Pi's native compaction when economic and context-window checks allow it. The extension operates through Pi's public APIs without patching or vendoring Pi, leaves the mechanisms disabled by default, and preserves original observations locally. It requires Node.js 22.19 or newer and is released under the MIT License.
Supabase is a hosted Postgres development platform for building production web, mobile, and AI applications, built from open-source components and available for self-hosting or local development. It provides a managed Postgres database, authentication and authorization, Row Level Security, auto-generated REST and GraphQL APIs, realtime database-change subscriptions over WebSockets, Edge Functions and serverless database functions, file storage, and an AI vector and embeddings toolkit. Its architecture combines PostgreSQL with services including PostgREST for REST APIs, GoTrue for JWT-based authentication and session management, Realtime for converting PostgreSQL replication changes to JSON broadcasts, Storage for S3-compatible file management with Postgres permissions, and pg_graphql for GraphQL access. It also provides dashboard tools such as table and SQL editors, policy management, deployment workflows, and official client libraries for JavaScript, Flutter, Swift, and Python, with additional community libraries for other languages.
typesafe-computer-use is an open-source macOS computer-use tool that drives a Mac toward a plain-English goal. It captures the frontmost window, reads text with OCR, collects labelled controls from the macOS accessibility tree, parses dates, and sends a structured list of possible actions to the TypeSafe decision model instead of sending screenshots to a frontier model for every step. The loop chooses actions such as clicking, scrolling, browser navigation, waiting, or pressing keys; a writing model is called only for free-text entry, proposed URLs, and the final answer, while passwords are never typed. It supports dry runs, confidence and step limits, annotated captures, detailed action and probability logs, and offline replay of saved screenshots. The repository targets macOS 14 or newer with Python 3.12 or newer, requires a TypeSafe API key, and is licensed under the MIT License.
Video Gen Client is an asynchronous Python client for video-generation APIs from Seedance, Kling, MiniMax/Hailuo, and Wan. It provides a shared interface for text-to-video and image-to-video generation, with provider and model discovery, task-status polling, and MP4 output. The project includes a command-line interface for listing providers and models, submitting generation jobs, checking task status, and serving the application. Its bundled FastAPI web UI supports provider and model selection, prompts, negative prompts, duration, aspect ratio, resolution, live task polling, video preview, MP4 download, and in-memory task history; it is served from the Python package without an npm, Node.js, or frontend build step. Provider adapters depend on external APIs and model IDs that may change independently of the client and therefore may require updates. The project is licensed under the MIT License.
Zetta is an open-source closed-loop embodied harness for self-evolving physical intelligence. It keeps a base policy frozen while evolving code-based runtime critics and recovery skills when rollouts fail. Its evolution protocol clusters failed trajectories, uses a diagnostic stage to identify an observable causal failure mechanism, generates Critic-Recovery candidates under fixed tool schemas, and checks them through shadow replay, paired same-seed evaluation, and held-out seeds. Candidates are rejected or promoted based on these gates; the critic may propose recovery actions, a higher-level role approves them, and a bounded recovery actor executes only approved programs. The repository provides campaign management, rollout workers and environment backends for systems including LIBERO, RoboCasa, RoboTwin, and Genie Sim, along with manifests, tool catalogs, audit records, deployment helpers, and contract tests. It supports Python 3.10–3.12, while simulator assets, model weights, credentials, and host runtime files remain outside the repository.
Searchable transcript of Top Dev Tool Projects : Claude Code, Supabase, Anki, rustfs, GitDiagram & OpenSpec — ManuAGI - AutoGPT Tutorials (14:37). Search for a phrase, then click its timestamp to jump straight to that moment in the video.
Captions sourced from the original video on YouTube, published by ManuAGI - AutoGPT Tutorials. The video, its captions and all related intellectual property remain the property of their respective owners; AINotes claims no ownership. Provided for research, accessibility and search — see the Transcript Notice and Copyright Policy.
00:00 Every week, developers release powerful open-source dev tools, and this weekly DevTool project update video brings them together in one place. Top trending open-source and best DevTool projects this week cover coding agents, agent skills, back-end platforms, object storage, and repo diagramming. You'll discover useful and trending developer tools you can start using right away.
00:22 Explained fast and to the point. Without wasting time, let's get started. Before we jump into today's project updates, here's a quick announcement for everyone, we've launched a brand new YouTube channel called AI Agent Studio, dedicated entirely to AI Agent projects, tutorials, and tools. So, if you're interested in staying up to date with the latest AI Agent open source projects, learning how to build your own agents, or exploring cuttingedge agent frameworks, make sure to check it out.
00:53 Subscribe now to get weekly videos, in-depth guides, and real-time project breakdowns. The link is right there in the description. Don't miss it. All right, let's get into today's video. >> Project number one, security audit skill. Turn your coding agent into a security auditor. Security audit skill is a Cloudflare skill that makes a coding agent run a structured security audit of a codebase.
01:15 It works in six phases. mapping the code, hunting for issues with isolated sub aents, then having a separate agent try to disprove each finding. So the checker is never the finder. Results are written as machine readable records split into confirmed needs validation and rejected with severity based on real impact. It runs risky target code only inside a sandbox and repeated runs widen coverage released under MIT.
01:41 Install it and audit your codebase. This one appeared in an earlier video. Video two, item two, but you're numbering it fresh here with the new short format. So, here's the 4080word version. Project number two, agent skills. Production engineering workflows for coding agents. Agent Skills is a pack of 25 engineering skills that make an AI coding agent follow the discipline workflow senior engineers use across the whole life cycle from spec to ship.
02:09 Nine SL commands cover defining, planning, building, testing, reviewing, and shipping. And skills also activate automatically by task. Each encodes steps, verification gates, and anti-shortcut rules, and installs into claude code, cursor, codecs, and many other agents released under MIT. Install it and start with a spec. Project number three, Claude code, Anthropic's agentic coding tool in your terminal.
02:35 Claude Code is Anthropic's agentic coding tool that lives in your terminal, understands your codebase, and works through plain natural language commands. It runs routine tasks, explains complex code, and handles Git workflows for you, and can be used in the terminal, in your IDE, or by tagging it on GitHub. It installs with a single script on Mac OS, Linux, or Windows.
02:58 Extends through plugins that add custom commands and agents, and connects external tools. Start it in code by describing what you want. Project number four, Bend. A language that blocks AI mistakes with proof. Bend is a fast parallel programming language built so you can trust AI written code without reading it. You declare rules your app must never break in a laws file.
03:20 And the compiler demands a mathematical proof that every edit still holds them. So an AI mistake that would violate a rule is blocked rather than merged. It runs code across CPU and GPU cores automatically with Python-like syntax and dependent types. Apache 2.0 and still young. Install it and tell your agent to use bend. Project number five, NAS A open speech model that generates and edits by instruction.
03:47 A a is an open-source 1.5 billion parameter foundation model from Tencent for generating and editing speech through plain natural language instructions. One interface covers many tasks. Zero shot and description-based text to speech, editing what is said, and changing pitch, speed, volume, emotion, tambber, or accent. Plus, speech enhancement and separating one speaker from a mix.
04:11 It runs on Nvidia GPUs through a command line, a web demo, comfy UI, or Python with a fast distilled variant. Released under MIT, install it and speak text in any voice. Project number six, SOOLP. Cut token waste in the PI coding agent. SOOLP is an NVIDIA extension for the PI coding agent that reduces token cost on longunning tasks without cutting the actual work.
04:36 It adds four opt-in efficiency mechanisms, running a follow-up validation command inside the same edit call, turning repeated large results into recallable handles, compacting long logs into receipts only when quotes match the source, and compacting finished plan steps out of context. It installs on an unmodified PI and preserves all evidence. Released under MIT, install it and trim your agents token bill.
05:01 Project number seven, CADX Astro Orchestrator. Two model Cadex setup that plans and builds. CADX Astro Orchestrator is a ready-made CADX configuration that splits work between two models. A stronger Astro model acts as the orchestrator that plans and integrates while a Luna model runs the specialized sub aents that do the actual work. It ships RO files for explorer, worker, tester, reviewer, and researcher agents, each pinned to Luna, plus a skill that Cadex can invoke to run the whole flow.
05:32 You copy the config into your repo or global setup and tune reasoning effort and concurrency. Install it and orchestrate your first task. Project number eight, Zetta, self-evolving harness for robot policies. Zetta is a closed loop embodied harness that lets a robot policy improve itself online while the base model stays frozen. Instead of retraining, it writes codebased runtime critics and recovery skills using three time scale separated loops that govern action frequency.
06:00 propose critic recovery fixes from failed rollouts and promote them only after passing same seed and held out validation gates. It targets simulated robotics benchmarks like Libro and Robocaza with learn skills transferring zero shot Python-based research project. Set up a campaign and evolve your policy. Project number nine, Rune GPU accelerated IDE and terminal multiplexer.
06:24 Rune is a fast GPU accelerated development environment that combines a fullfeatured IDE with a terminal multiplexer built as a native application rather than on Electron. It suits both manual coding and driving AI agents with production-grade language intelligence builtin and a flexible screen layout of workspace slots, terminals, tabs, and windows.
06:45 Every instance forms an endto-end encrypted peer network so you can connect between your machines and continue work from anywhere. and it extends through official or git hosted plugins written in Go GPL 3.0. Install it and organize your workspace. Project number 10, Agent Skills. Senior engineer discipline for your coding agent. Agent Skills is a pack of 25 reusable skills that push an AI coding agent to work like a senior engineer instead of taking the shortest path.
07:16 It spans the full life cycle, defining, planning, building, testing, reviewing, and shipping, reached through nine/comands or triggered automatically by the task. Each skill is a step-by-step workflow with verification gates and tables that rebutt the excuses agents use to skip steps, and it installs into clawed code, cursor, codecs, and dozens more.
07:37 MIT licensed. Install it and run your first spec. Project number 11. Open spec. Agree on a spec before the AI writes code. OpenSpec is a lightweight specdriven development tool that gets you and your AI coding assistant to agree on what to build before any code is written. Through SL commands, you explore an idea, then generate a change folder holding a proposal, plain markdown requirements with concrete scenarios, a design, and a task checklist which you review before implementation.
08:07 It archives completed changes and update specs, works across 30 plus AI tools, and offers shared crossreo planning stores for teams. Node-based MIT licensed. Install it and propose your first change. Project number 12, Anki spaced repetition flashcards that time your reviews. Anki is a spaced repetition flashcard program that helps you memorize and retain information efficiently.
08:32 Instead of reviewing everything equally, it schedules each card based on how well you know it, showing material just before you would forget it, so you study less and remember more. This repository holds the source for the computer version built with Rust, Python, and TypeScript. And it syncs with companion mobile and web apps, open- source, build it or download it and study your first deck.
08:54 Project number 13, Rust FS, S3 compatible object storage built in Rust. Rust FS is an open-source distributed object storage system written in Rust. Positioned as a memory safe Apache 2.0 licensed alternative to Mino. It offers broad S3 API compatibility so existing S3 clients and tools work with it and it can migrate from or run alongside other S3 compatible platforms like Mino and Seph.
09:22 It supports versioning, bucket replication, event notifications, bit rot protection, multi-tenency, and the OpenStack Swift and Keystone protocols running single node or distributed. You install it with a script, Docker, Helm or Nyx, and manage it through a web console, deploy it, and create your first bucket. Project number 14, Dissipostgress back-end platform, an open Firebase alternative.
09:47 Dissip is an open-source development platform that gives you a dedicated Postgress database with the surrounding features to build web, mobile, and AI apps positioned as an open alternative to Firebase. Around the database, it adds authentication, autogenerated REST, and GraphQL APIs, real-time subscriptions over websockets, file storage, edge functions, and a vector toolkit for embeddings, all managed through a dashboard.
10:12 It is assembled from open-source tools, offers client libraries in many languages, and runs hosted or self-hosted, Apache 2.0, sign up or self-host, and create your first project. Project number 15, Get Diagram. Turn any GitHub repo into an architecture diagram. Git diagram turns any public or private GitHub repository into an interactive architecture diagram in seconds, so you can grasp a codebase's structure instead of reading through folders.
10:40 It reads the repository tree and readme has an AI model write a plain English explanation. Then generates a validated system level graph where clicking a component opens its real file on GitHub. You can swap hub for diagram in any repo URL to open it. Use a local token for private repos and export the diagram as mermaid or PNG. Built with Nex.js MIT licensed.
11:02 Paste a repo and see its diagram. Project number 16 bird view. Map the architecture before an AI edits it. Bird view is a skill that makes an AI coding agent map a project's architecture before it changes any code. So you see which modules a change will touch before it happens. From an architecture file and the agents declared activity, it renders a standalone interactive HTML view showing modules, ownership, evidence, and plan scope with architecture changes and comparison views.
11:34 It validates both inputs against JSON schemas and works with any agent through agents.md guidance. Nodebased MIT licensed install it and map your project. First project number 17, videogen client. One Python client for four video generation services. Video genen client is an async Python wrapper that gives you one interface to several AI video generation providers covering seedance, cling, miniax/ Halu, and WAN across many model versions.
12:05 Instead of learning each provider's API, you generate texttovideo or imagetovideo clips through one client, adding your own provider keys. It offers a command line tool to list providers and models, start jobs and poll status, plus a local fast API web UI with model selection prompts and MP4 download. And it includes an MCP server MIT licensed. Install it and generate your first clip.
12:29 Project number 18, Agentverse OS. Personal cloud OS for you and your agents. Agentverse OS is a personal cloud operating system that runs on a single Iuntu server and gives one developer and their AI agents a full environment reached entirely through the browser. One command installs it. Then you get a Window desktop that adapts to laptop, tablet or phone.
12:51 Isolated per project workspaces with browser VS Code plus cloud code and codecs. A store of 944 self-hosted apps, backups and updates. Access goes only through tail scale. Nothing exposed to the internet. Rust Core Spelt UI Apache 2.0 Alpha. Install it and open your cloud in the browser. Project number 19. Page Mascot. An interactive character that follows your cursor.
13:18 Page Mascot is a small web component that puts an interactive character on your page that watches the cursor and blinks when you poke it. Each mascot is two sprite sheets, nine head directions, and nine expressions. and the pointers angle picks the direction while a click plays a reaction. You install it from npm and point it at two image sheets, choosing from ready-made characters or drawing your own through an installable agent skill that generates and aligns the sheets.
13:44 It respects reduced motion and fine pointer settings. MIT licensed. Install it and add a mascot to your page. Project number 20. Typesafe.computer use. Drive your Mac cheaply without screenshots to a big model. Typesafe.computer Computer use drives a Mac toward a plain English goal for about a 50th of a cent per step without sending screenshots to a large model.
14:08 It reads the screen deterministically with OCR and accessibility data. Then asks a small type-S safe classifier which action comes next from a fixed list clicking, scrolling or opening a site and only calls a writing model when a text field genuinely needs free text. Each step logs an annotated screenshot and payload for replay. Python, Mac OS only, MIT licensed. Install it and give your Mac a goal. Thanks for watching.