lab

Use when creating, editing, validating, testing, running, explaining, or removing Labby Code Mode snippets; when a user wants a reusable workflow made from gateway MCP tools; or when building schema-backed snippets from upstream tool ids, JSON schemas, params, inputs, defaults, artifacts, and MCP/CLI snippet actions.

Use when creating, editing, validating, testing, running, explaining, or removing Labby Code Mode snippets; when a user wants a reusable workflow made from gateway MCP tools; or when building schema-backed snippets from upstream tool ids, JSON schemas, params, inputs, defaults, artifacts, and MCP/CLI snippet actions.

This skill should be used when the user wants to automate an iOS Simulator, drive an Android device or emulator via MCP tools, test on multiple devices simultaneously, run a visual regression baseline on mobile, audit accessibility on a real device, release an app to the Play Store, or use the claude-in-mobile CLI or MCP server. Does not apply for structured end-to-end test reports (use android-app-testing or desktop-app-testing for those).

Interpret and use `mcpjam` probe, doctor, OAuth, apps conformance, tools, resources, and prompts output conservatively against MCP 2025-11-25. Use when interacting with MCP servers, executing tools, triaging findings, performing security reviews, deciding whether a CLI finding is real or overstated, or turning inspection output into an engineer-facing report with severity and confidence.

Audit the current Git checkout, open worktrees, local branches, stale or merged cleanup candidates, merge readiness, conflicts, PR/CI/test state, blockers, and safest merge order. Use when the user asks for repo status, branch/worktree cleanup candidates, stale branch review, conflict investigation, merge readiness, or what must be done before open branches can merge.

This skill should be used when the user mentions labby, the labby CLI, the Lab gateway, or any Lab operator surface. Triggers include: "run labby doctor", "check labby health", "start the labby MCP server", "configure ~/.lab/.env", "search upstream MCP tools with Code Mode", "use labby gateway to import servers", "manage the Labby marketplace", "reload the gateway", or any request to run labby CLI commands, inspect gateway upstreams, or dispatch an action against a Lab service.

Use when invoking upstream MCP tools through the Lab gateway Code Mode surface — i.e. when the exposed MCP tools are `search` and `execute`, and real upstream tools are hidden behind the gateway. Trigger on "search", "execute", "gateway tools", "find a tool for...", "call X through the gateway", "what tools do I have", or whenever `list_tools` shows the synthetic `search` + `execute` pair. Not for designing or implementing new MCP tools, building MCP servers, or generic "what can you do" questions unrelated to the gateway.

This skill should be used when the user asks to save a snippet, store code, create a snippet, search snippets, find a snippet, share a snippet, list snippets, delete a snippet, add to ByteStash, or mentions ByteStash or snippet management.

This skill should be used when the user mentions Immich or asks to browse their photo library, search for photos, find pictures from a trip or date range, list albums, check server stats, or view storage usage. Triggers include: "find photos of", "show my albums", "search Immich", "how many photos do I have", "check my Immich server", or any question about a self-hosted photo library.

This skill should be used when the user asks to save a bookmark, add or save a link, search bookmarks, list bookmarks, find saved links, tag a bookmark, archive a bookmark, check if a URL is already saved, list tags, or create a bundle. It also applies when the user mentions Linkding by name or asks about their bookmark library.

This skill should be used when the user asks to save a note, create a memo, search memos, find notes about something, add a note, capture a thought, save something to their note hub, or mentions the Memos service. Does not apply when the user says 'remember this' without specifying Memos — that may route to the mnem memory system instead.

This skill should be used when the user wants to interact with their Navidrome self-hosted music server — checking server status, browsing artists or albums, searching the music library, listing playlists, viewing now-playing or recently played tracks, or accessing starred items. Triggers include: "what's playing on Navidrome", "show my playlists", "search for an artist", "recent albums", "Navidrome status".

This skill should be used when managing calendars and contacts on a self-hosted Radicale CalDAV/CardDAV server. Use when the user asks to "list my calendar", "what's on my calendar this week", "show me my events", "when is my next event", "add to my calendar", "create an event", "schedule a meeting", "schedule an event", "delete an event", "cancel event", "remove event", "find a contact", "what's someone's email", "search my contacts", "look up [person] in contacts", "add a contact", "save contact", "save someone's phone number", or mentions Radicale, CalDAV, CardDAV, calendar events, or contact management operations.

Manage documents in Paperless-ngx document management system. Use when the user asks to "upload document", "search paperless", "find document", "add to paperless", "tag document", "manage correspondents", "organize documents", "archive document", "export document", "delete document", or mentions Paperless-ngx, document management, OCR, or paperless office.

Use when the user wants to live-test a built Windows desktop application (.exe) end-to-end inside the agent-os Windows 11 VM and get a works/doesn't-work + UI/UX report — launching the real binary and driving it, not writing test code. Triggers: "test my Windows app", "QA this .exe", "run my desktop build on agent-os and tell me what breaks", "click through the app", "review the desktop app's UX", "does my exe work", "test the built binary in the Windows VM". Gets the .exe into the VM, launches it, enumerates controls from the UI Automation tree, drives every feature (click/type by element), watches for crashes/hangs/error dialogs, captures screenshots + control-tree dumps, and emits a structured report. Drives the agent-os VM via the Windows-MCP gateway. Sibling of web-app-testing and android-app-testing (shared report format). Does NOT fire for: building/coding a desktop app, the user's personal Windows on steamy (use nircmd), or general agent-os VM driving (use agent-os).

Use when the user wants to live-test a WEB app/site end-to-end in a real browser and get a works/doesn't-work + UI/UX report — not just write Playwright code. Triggers: "test my web app", "QA this site", "run E2E on the web app", "click through every feature and tell me what breaks", "review the UX of this web app", "does my web app work", "test the deployed site". Drives a real Chrome via Playwright over CDP (default http://127.0.0.1:9222), enumerates features from the DOM/ARIA tree, exercises each, captures screenshots + console/page/network errors, and emits a structured report. Sibling of android-app-testing and desktop-app-testing (shared report format). Does NOT fire for: writing a one-off Playwright script (use webwright), pure web design/build with no testing, or backend/API-only testing.

This skill should be used when the user wants to drive, control, screenshot, run PowerShell on, install software on, or interact with the agent-os Windows 11 sandbox VM. Triggers include: "run X on agent-os", "screenshot agent-os", "PowerShell on agent-os", "spin up agent-os", "drive the windows VM", "the windows sandbox", or "winbox". Also applies for web-dev browser verification against the agent-os Chrome endpoint. Does not apply for the user's personal Windows on steamy (use the nircmd skill). The sandbox is Claude's alone — install software, change settings, run shells freely without asking for confirmation.

Bridge for viewing the user's own desktop when they're SSH'd in and can't paste images. Three modes — (1) fetch an existing screenshot from the user's screenshots dir, (2) take a fresh full-screen / region / active-window capture via ShareX CLI on a Windows target, (3) fall back to a PowerShell-only capture if ShareX isn't installed. Use when the user says "check my screenshot", "look at my screen", "latest screen", "take a screenshot of my desktop", "capture my screen", "show me my desktop", "screenshot this window", "grab a region of my screen", or similar references to seeing their own machine. Defaults target the user's Win11 box via `ssh steamy-wsl`; override `SCREENS_HOST` (and friends) if pointing at a different host. **For Chrome tab screenshots (any window state, including minimized), use the `chrome` skill instead — this skill is for desktop pixels.**

This skill should be used when implementing an ACP agent or extending the lab ACP runtime in Rust — including using Client.builder()/ByteStreams/attach_session to connect to ACP providers, calling session_config_options() to discover models and config options, switching models via SetSessionConfigOptionRequest, implementing the Agent trait for a new ACP provider (stdio agent), handling session/prompt or session/update wire messages, or debugging JSON-RPC 2.0 stdio transport issues. Also applies when working on crates/lab/src/acp/runtime.rs, the codex-acp reference implementation, or authoring bidirectional stdio agents for Zed or VS Code.

This skill should be used when the user asks about Jellyfin media server. Triggers include: "check Jellyfin", "Jellyfin library", "who's watching on Jellyfin", "active Jellyfin sessions", "add a Jellyfin user", "Jellyfin metadata", "Jellyfin transcoding", "Jellyfin health", "Jellyfin scheduled tasks", "Jellyfin plugins", or any mention of Jellyfin media server management.

This skill should be used when the user wants to request movies or TV shows via Overseerr, monitor or manage media requests, or check request status. Triggers include: "request a movie", "request a TV show", "add to Overseerr", "check request status", "pending requests", "is my request done", "Overseerr status", or any mention of Overseerr media requesting.

This skill should be used when the user wants to interact with their Plex Media Server. Triggers include: "check Plex", "search Plex", "what's on Plex", "what's playing on Plex", "who's watching", "Plex sessions", "active streams", "Plex library", "browse movies", "browse TV shows", "recently added", "on deck", "continue watching", "Plex status", or any mention of Plex Media Server.

This skill should be used when the user asks about torrents, downloading, seeding, or the qBittorrent client. Triggers include: "what's downloading", "list torrents", "add a torrent", "pause/resume/delete torrent", "torrent speed", "download queue", "qbit", "qBittorrent stats", "check download status", or any mention of managing a torrent client.

This skill should be used when the user wants to manage TV shows in Sonarr. Triggers include: "add a TV show", "add to Sonarr", "search Sonarr", "find a series", "remove a show", "delete show", "check if show exists", "is [show] in my library", "what's airing this week", "upcoming episodes", "Sonarr library", or any general mention of Sonarr or TV show library management.

This skill should be used when the user asks about Plex watch statistics, current streams, who is watching Plex, active sessions, playback history, most-watched content, user activity, library stats, stream analytics, or anything related to Tautulli monitoring and analytics.

This skill should be used when the user mentions operating on a named remote host that has a Plexus `remotes/<host>/REMOTE.md` profile, such as squirts, dookie, steamy, or another homelab device. Before taking action, load durable host memory and live context with `remote-context.py <host>`. This is the host-scoped equivalent of CLAUDE.md/AGENTS.md for remote machines.

This skill should be used when the user wants to validate an MCP-UI or MCP Apps implementation — checking widget rendering in the Inspector, verifying ui:// resource contracts, testing structuredContent, or running MCP Apps conformance checks. Triggers include: "why isn't my MCP widget rendering", "test my MCP-UI implementation", "check if my tool passes MCP Apps conformance", "verify the ui:// resource contract", "debug my Inspector view". Does not apply for general MCP tool smoke-testing (use mcporter for that).

Use when the user mentions mcporter, says "test an MCP server", "smoke-test these tools", "automate MCP testing", "call a tool from the shell", "list MCP tools", "exercise the gateway tools", or asks for a script that hits MCP endpoints. Covers using mcporter to discover, inspect, and call MCP servers from the shell, and to write repeatable regression or smoke-test scripts. Not for designing new MCP servers, writing server-side handlers, or generic API testing unrelated to MCP.

This skill covers building, modifying, and debugging MCP (Model Context Protocol) servers and clients in Rust using the rmcp crate. It applies when the codebase imports `rmcp`; when defining tools, resources, or prompts with `#[tool]`, `#[tool_router]`, or `#[prompt_router]`; when choosing or wiring transports (stdio, TCP, Unix socket, HTTP Streamable); when implementing `ServerHandler` or `ClientHandler`; when sending progress notifications; or when a user asks to "add an MCP tool", "create an MCP server", "connect to an MCP server", or "implement a handler".

Use when the user wants to live-test an Android APK end-to-end on an emulator/device and get a works/doesn't-work + UI/UX report — driving the real app, not writing test code. Triggers: "test my Android app", "QA this APK", "run the app on the emulator and tell me what breaks", "click through every screen", "review the app's UX", "does my APK work", "test the built APK". Installs/launches the APK, enumerates screens from the accessibility tree, exercises each feature via adb (tap/type/swipe/launch), watches logcat for crashes/ANRs, captures screenshots + UI dumps, and emits a structured report. Primary path is direct local adb; optional richer path via claude-in-mobile. Sibling of web-app-testing and desktop-app-testing (shared report format). Does NOT fire for: building/coding an Android app (use claude-android-ninja / jetpack-compose-expert), iOS, or unit tests.

Use whenever the user wants to add or scaffold a new reverse proxy config for SWAG — the LinuxServer.io SWAG proxy on `squirts` (`/mnt/appdata/swag/nginx/proxy-confs/`) that fronts every `*.tootie.tv` subdomain. Triggers: "create a swag config", "add a swag proxy for X", "add X to swag", "make a subdomain config", "expose X on tootie.tv", "add reverse proxy for X", "new tootie.tv subdomain", "proxy X through swag", "scaffold a swag entry", "wire up a SWAG mcp config". Always prefer the `swag-mcp` MCP server (registered as `swag` at `https://swag.tootie.tv/mcp`, exposes one action-routed tool: `swag` with actions list/create/view/edit/update/remove/logs/backups/health_check) over hand-writing files — it validates inputs, renders the canonical Jinja2 template, writes the file with the correct name, manages backups, and exposes its own health_check. Hand-write only when swag-mcp is unreachable. Does NOT trigger for nginx work outside this homelab.

Use when the user wants to live-test a built Windows desktop application (.exe) end-to-end inside the agent-os Windows 11 VM and get a works/doesn't-work + UI/UX report — launching the real binary and driving it, not writing test code. Triggers: "test my Windows app", "QA this .exe", "run my desktop build on agent-os and tell me what breaks", "click through the app", "review the desktop app's UX", "does my exe work", "test the built binary in the Windows VM". Gets the .exe into the VM, launches it, enumerates controls from the UI Automation tree, drives every feature (click/type by element), watches for crashes/hangs/error dialogs, captures screenshots + control-tree dumps, and emits a structured report. Drives the agent-os VM via the Windows-MCP gateway. Sibling of web-app-testing and android-app-testing (shared report format). Does NOT fire for: building/coding a desktop app, the user's personal Windows on steamy (use nircmd), or general agent-os VM driving (use agent-os).

Create the save-to-md session doc before staging, git add all, commit with Claude co-authorship trailer, and push to current/new feature branch — including project version bump and changelog update when applicable. Use when the user says "quick push", "push my changes", "commit and push", "ship this", "push to a new branch", or any request to wrap up local work and get it on the remote. Accepts optional `--no-bump` argument to skip the version bump.

Use when the user asks to drive or test Claude's reserved agent-os Windows 11 sandbox VM via Windows-MCP, PowerShell, screenshots, desktop apps, installers, registry, filesystem, or noVNC. Triggers include agent-os, the agent-os VM or desktop, windows sandbox, winbox, run on agent-os, screenshot agent-os, or PowerShell on agent-os. Prefer webwright for generic web verification; use agent-os only when the task depends on the Windows sandbox or desktop. Do not use for the user's personal Windows machines such as steamy or steamy-wsl.

This skill should be used when the user mentions NotebookLM, says /notebooklm, or asks to create a podcast, generate an audio overview, make a quiz, summarize URLs, add sources to NotebookLM, generate flashcards, create a mind map, make an infographic, or download generated content. Covers full programmatic access to Google NotebookLM including features unavailable in the web UI.

This skill should be used when the user wants to install, set up, repair, or initialize the Plexus remote-device memory plugin; when they ask where Plexus stores persistent REMOTE.md host profiles; or when host profiles are missing and need to be seeded from bundled defaults.

This skill should be used when the user asks to verify, audit, or report on CLI tools required by their agent skills. Triggers include: "check whether skill CLIs are installed", "are all my skill tools available", "which commands do my skills need", "find missing skill dependencies", "compare enabled vs disabled skill CLIs", or "produce a report of missing agent tools" across Claude, Codex, and Copilot surfaces.

This skill should be used when the user wants to inspect or control their own Chrome browser session over CDP. Triggers include: "grab my chrome tab", "show me my tabs", "screenshot my [site] tab", "what's open in my chrome", "eval this in my browser", "cookies from my chrome", "navigate my chrome to", "what's the page console showing", "check my chrome network requests". Does not apply for generic browser automation (use webwright) or the agent-os Chrome endpoint (use agent-os).

This skill should be used when the user asks about AdGuard Home status, DNS filtering, or blocked domains. Triggers include: "is adguard running", "how many queries did it block today", "show me recent DNS queries", "search the DNS log for a domain", "is example.com blocked", "AdGuard stats", or any question about DNS-level ad blocking or network filtering.

This skill should be used when the user asks to search indexers, find a release, check indexer status, list indexers, sync indexers to Sonarr or Radarr, connect Prowlarr to another app, or mentions Prowlarr or indexer management.