Zero-dependency static + sandbox smoke detector for DeepSeek Harness (dsh) plugins: package-structure gates (R), cordis contract scans (K), keyless-headless sandbox smoke (D), and ecosystem-listing checks (CC).

Desktop notifications for DSH Desktop that actually speak: Windows TTS out of the box (zero download), VOICEVOX voice packs optional, a selectable system-sound fallback, and an idle gate that waits for running subagents before saying the turn is done. Inc

Interactive isometric line figures for React: rounded boxes, faces you can draw on, keys you can press, and optional synthesized sound. Plus a Claude Code skill that draws them.

Legacy: build a Bazel iOS app with `lim xcode rbe` on a dedicated Xcode sandbox and install it on a remote iOS simulator attached to that sandbox. Use only when the workspace already runs on `lim xcode rbe` (its `.limrun/` came from `lim xcode rbe`) or the user explicitly asks for it, or when a `lim

Your mailboxes in ChatGPT and Claude: Gmail, iCloud, Fastmail, any IMAP. Passwords stay yours.

Build an Android app on a remote Gradle sandbox with `lim gradle build` instead of local Gradle or Android Studio, from any environment (Linux, Windows, macOS, VM, container). Use when the user wants to build an APK or AAB, sign a release with an upload key, prepare a Play Store publish, inspect bui

Music video player for DeepSeek Harness Desktop: text, Canvas2D and WebGL2 3D scenes synced to your own audio. Sandboxed workshop packs can include licensed bilingual lyrics, timing and visual assets; only music is user-provided. Plugin code MIT; no song

Use MCP 2026-07-28 discovery or 2025-11-25 initialize to list and execute PubFi routes.

Execute Python code in a safe sandboxed environment via [inference.sh](https://inference.sh). Pre-installed: NumPy, Pandas, Matplotlib, requests, BeautifulSoup, Selenium, Playwright, MoviePy, Pillow, OpenCV, trimesh, and 100+ more libraries. Use for: data processing, web scraping, image manipulation

A CLI that turns a goal into a pull request - and a sandbox for testing concurrent AI coding agents on the Mozaik runtime.

InnoShop is an AI-powered open source e-commerce system built on Laravel 13. It supports multiple AI models and native MCP (Model Context Protocol) for intelligent automation, with multi-language and multi-currency out of the box. Modular architecture and flexible plugin/theme system make it highly adaptable to diverse business needs.

Execute Python code in a safe sandboxed environment via [inference.sh](https://inference.sh). Pre-installed: NumPy, Pandas, Matplotlib, requests, BeautifulSoup, Selenium, Playwright, MoviePy, Pillow, OpenCV, trimesh, and 100+ more libraries. Use for: data processing, web scraping, image manipulation

Enables Claude to brief self-hosted GrokBot bots through their webhook routines and then collect their replies in the same thread, with tools for waiting on replies, checking the inbox, and reading conversations. It includes a self-hosted dashboard for registering bots, managing encrypted webhook keys and tokens, and connecting Claude apps over OAuth or API tokens.

Enables AI agents to make policy-scoped payments with spending caps, recipient allowlists, approval thresholds, instant revocation, and an append-only audit trail in a sandbox wallet.

Enables agents to reach pump.fun's full API surface — official developer routes, bonding-curve reads, Solana RPC, and unsigned trade building — by writing JavaScript against one merged OpenAPI document, using just two tools (search and execute). Model-authored code runs in a network-less sandbox, so multi-step questions resolve in a single round trip while credentials stay on the host Worker.

Turns any OpenAPI spec into a single-tool MCP server where the model writes JavaScript against a typed api object in a QuickJS sandbox, so multi-step tasks run in one round trip and only the final result reaches context. Credentials, rate limiting, retries, write gating and audit logging stay on the host, outside the sandbox.

Enables MCP clients to create and control persistent, PTY-backed terminal sessions on macOS: launch commands and interactive programs, read output incrementally via byte cursors, send keystrokes, wait on conditions, and detect running/awaiting-input/exited states. Sessions run inside a Seatbelt sandbox with command policies, human approvals, secret redaction, and hash-chained audit logging.

Enables two AI agents to share a persistent task mailbox over remote MCP, so one can leave work, the other can claim and complete it, and results are written back with optional webhook wake-ups.

Enables an LLM to orchestrate multiple downstream MCP tools inside one isolated sandbox in a single turn, filtering intermediate results in memory and returning only the final output to cut context tokens.

Enables AI clients to securely access and manage IMAP mailboxes through OAuth-protected tools for searching, reading, organizing, flagging, moving, and deleting emails or folders, with optional write permissions.

Enables MCP clients like Claude Code, Codex, Cursor, and Claude Desktop to search, read, and download attachments from NetEase mailboxes (163/126/yeah.net/188), wait for verification codes, and create drafts or send and move messages only after a two-stage preview and confirmation, with readonly, draft, and full permission modes.

Enables Claude to work across several Google accounts at once, searching every Gmail inbox, reading and drafting email, browsing Drive, and checking or scheduling across multiple Calendars in a single conversation. It supports tags for grouping accounts and cross-account tasks, and only creates drafts, files, and events rather than sending anything without you.

Enables AI agents to browse and interact with the noymanga.com community through MCP tools and resources, including reading posts, rankings, and inbox messages; uploading images; creating posts and comments; liking posts; checking in; and managing badges.

Enables coding agents on different machines to exchange end-to-end encrypted messages through a shared, self-hosted relay, exposing inboxes, room history, and notifications to Claude Code, Codex, or any MCP client. Peers' messages arrive automatically as clearly framed untrusted data, so collaborators can coordinate without granting tool, file, or command access.