Recently updated open-source projects

Start an Electron desktop project with a Vite and React TypeScript setup that is ready for development, testing, and packaging. The template includes main-process and preload structure, debugger configuration, Vitest and Playwright tests, electron-builder packaging, native Node module support, and room for multiple windows. Its renderer adapter exposes compatibility helpers but does not enable Node integration by itself; direct Node access requires an explicit Electron setting and a careful security review.

Blitz is a native macOS tool that lets Claude Code, Codex, or another MCP client operate an iOS app's App Store Connect workflow. Agents can manage versions and pricing, configure in-app purchases or subscriptions, upload localized screenshots, work with TestFlight builds and testers, control simulators, and submit an app for review. It is aimed at iOS developers who want terminal-driven automation alongside a native graphical interface. Official release builds may send limited anonymous product telemetry, while source, fork, and debug builds keep it off unless configured.

Give the skill a URL to run either a basic or full SEO audit and receive an HTML report organized around concrete fixes. Deterministic Python checks cover items such as HTTP behavior, robots and sitemaps, metadata, canonicals, and structured data, while the agent handles semantic judgments like keyword intent, content quality, and expected page type.

yume is a native desktop frontend that launches the official Claude Code CLI rather than replacing its runtime. It adds tabs, split panes, task orchestration, background agents, visual diffs, rollback, and crash recovery while preserving existing authentication, subscriptions, MCP servers, hooks, and skills. It can also route supported workflows to other model providers through its own shim. Claude use still requires a Claude account, and the free demo limits the number of tabs, panes, and windows.

OpenCovibe wraps Claude Code and Codex in a local-first desktop interface with visual tool cards, syntax-highlighted diffs, project files, and replayable session history. Users can resume or fork runs and switch among multiple model providers without restarting the app. Data stays on the machine, although remote model APIs still need network access, and macOS is currently the most thoroughly tested platform.

Pasteboard Viewer helps developers verify exactly what their apps place on NSPasteboard or UIPasteboard. It refreshes content live and can preview text, RTF, images, and other items supported by Quick Look, with an option to keep the macOS window in front. It is a debugging utility rather than a clipboard history manager, and the current release requires macOS 26, iOS 26, or visionOS 26, with older macOS builds available separately.

This repository packages portable Codex skills for two practical workflows: auditing a fast-moving project before commit or deployment, and applying stricter reliability habits during coding tasks. Clone it, copy or symlink the desired skill folders into your agent’s skills directory, then invoke them by name. The current catalog centers on the wtf audit skill and the better-codex reliability overlay.

Forge is a local-first Claude Code persona engine with separate workflows for modeling yourself and approximating another person's voice and interaction patterns. A self persona can support reflection and multi-agent decision discussions, while a person persona can be used for memory-based role dialogue. The repository includes parsers for chats, diaries, and social material, plus local persona files, validation, versioning, and decision logs. It is explicitly intended for personal memory and emotional healing, not harassment, tracking, or privacy invasion.

Claude Code Trace turns session files from the local Claude projects directory into readable conversations. It supports searching by user message, expanding tool and MCP calls, inspecting timestamps and token counts, and following active sessions in real time. The viewer is available as a desktop app, web interface, or terminal UI and does not require sending traces to an external service.

Hermes Agent Idea Workflow structures the pre-build phase before any coding begins. It can capture an idea in a lightweight mode or guide a full interview that produces a design document, an optional UI brief, an implementation specification, and a single-file handoff for another agent. A review gate marks the package as pass, pass with changes, or fail before handing it to a Superpowers-based build process. The workflow itself does not implement the product and remains in beta field testing.

Memory Forge RS is a fully local desktop session manager for Claude Code, Codex CLI, OpenCode, Kiro, and Gemini CLI. It lets users edit earlier messages, remove irrelevant history, or inject missing context so a resumed session continues from the corrected record instead of starting over. Every change is tracked with a before-and-after audit log, and the app also provides search, favorites, archives, Markdown export, session aliases, and a reusable prompt library.

Use the desktop app during calls, streams, or bilingual meetings to capture system audio and microphone speech, then display live translations in a compact overlay. It supports one-way or two-way conversations, several translation engines, optional narration, and custom terminology. The app connects directly to providers with locally stored keys; its fully offline mode is limited to Apple Silicon and a narrower language set.

Safe Solana Builder is an AI coding skill for Claude or Cursor that applies a security-first workflow before Solana code is written. It selects Anchor, native Rust, or Pinocchio, assesses risk, and applies rules for account validation, PDAs, CPIs, arithmetic, Token-2022, and related failure modes. Each request can produce a full project scaffold, compilable program, test file, and security checklist; security edge-case tests are scaffolded with TODO bodies, positioning the skill as a first hardening layer before audit.

SwarmClaw is a self-hosted runtime and dashboard for operating individual agents, swarms, and orchestrators. It provides durable memory, schedules and heartbeats, delegation, background jobs, MCP tools, and connections to many model APIs and coding-agent CLIs. It is aimed at long-running workflows that need observation and operator control, rather than an agent library embedded in another application.

AEGIS is a desktop monitoring tool for coding and autonomous agents that observes activity outside the agent runtime rather than relying on an installed hook. It tracks processes, file access, network connections, recognized agent signatures, risk scores, and behavioral anomalies, with activity views and printable threat reports. The current release is monitor-only and does not actively block actions; Anthropic-based threat assessment is an optional analysis layer.

OmniSearch is a Windows desktop app that reads NTFS MFT and USN metadata for fast global search, advanced filtering, duplicate detection, and common file actions. After the initial scan, a journal watcher keeps the index updated incrementally. An optional Android companion pairs by QR code over the local network for browsing and transferring files. The desktop engine requires Windows 10 or 11, an NTFS volume, and elevated access to read volume metadata.

PLAID is a documentation-driven agent skill that guides founders through product discovery, validation, planning, design, launch, and roadmap execution. Structured conversations capture the user's background, audience, business intent, constraints, product direction, and desired experience, then produce artifacts such as a product idea, validation report, vision, PRD, roadmap, design specification, and go-to-market plan. It requires no additional runtime dependencies, but it is meant to run inside an AI coding environment that can load and follow the skill.

AI Factory configures a project for selected coding agents, installs relevant skills, sets up MCP integrations, and adds spec-driven workflows for exploration, planning, implementation, fixes, review, and quality gates. It supports many agent tools, including Claude Code, Codex, and Cursor. A skills-only installation is available, but it omits the CLI, automatic MCP configuration, and interactive setup.