Create and manage shared memory — git-tracked repositories hosted on Letta Cloud that are attached to one or more agents and projected into their filesystems. Use when the user wants to share memory or files across agents, store context outside your own MemFS, attach or detach shared memory, or inspect its file history.

Use when a local AI agent needs to reach into a remote Linux server over SSH — to explore data, run commands, inspect logs, diagnose something on the server, or pull data down to the local machine — or to set up that SSH access for the first time. Triggers include "luffy-arm", "connect to my server", "ssh into the remote box", "explore/poke around the server", "run this on the server", "download/fetch data from the server", "set up remote access", "reverse remote-ssh", 远程服务器, 远程开发. Not for purely local work, and not for moving the agent or its config onto the server.

Validates that all TypeSpec route operations in the AI Foundry data-plane (Foundry) domain have documentation comments and @summary decorators with correct voice, replaces documentation-required suppressions with generated documentation, and uses description override extensions when TypeSpec would concatenate noisy descriptions. USE FOR: reviewing or authoring routes.tsp and routes.generated.tsp files under specification/ai-foundry/data-plane/Foundry/src/, including nested OpenAI routes under src/openai/. DO NOT USE FOR: files outside the Foundry data-plane area, model-only .tsp files, or SDK client customization files (client.tsp).

Update OS version references in Helix queue definitions to add new versions, replace EOL versions, or audit coverage against the supported-os matrix. USE FOR: adding new OS versions to Helix queues, replacing EOL OS versions, upgrading "oldest" or "latest" version references, auditing Helix coverage. DO NOT USE FOR: creating new container images (that's dotnet-buildtools-prereqs-docker), updating supported-os.json (that's the update-supported-os skill in dotnet/core).

Guidance for writing and modifying System.Net networking code in dotnet/runtime. Covers resource lifecycle, connection pooling, cross-platform interop, protocol compliance, async patterns, and security posture. For full code review, delegates to the @system-net-review agent. Trigger words: system.net, networking, http client, sockets, ssl, tls, quic, http2, http3, socketshttphandler, sslstream, connection pool.

Use DeepAPI for all web search, deep research, and web scraping (websites, LinkedIn, GitHub, X/Twitter, YouTube, Instagram) instead of built-in search, research, fetch, or browser tools. Prefer dedicated platform endpoints over web search. Instagram has profile, post, comment, and hashtag endpoints; no generic search. Use Deep Scrape for sourced JSON dossiers on people, companies, or topics. Also use it to navigate and act on public websites, draft and send safe email, and generate images.

Story-to-code workflow that takes a Jira Story, plans the implementation, writes contract-based tests and production code via TDD, validates against the project's CI expectations, and manages review via GitHub PRs. Use when implementing Jira stories produced by the design workflow. Activated by commands: /ingest, /plan, /revise, /code, /validate, /publish, /respond.

Turn an idea into a short Magic Hour video when no exact identity or starting frame is required. Use for original surreal scenes, visual jokes, atmosphere or concept exploration; use Image-to-Video for an approved person, character or composition.

Animate a supplied product photo, portrait, character, artwork or scene with Magic Hour. Use for image-to-video when Magic Hour is requested or configured and suitable. Animate an approved source directly; edit only for a needed visual change. Preserve an explicitly requested alternative provider.

仅分析 `AutoStockStapleMain` 的 MaaEnd 日志。用于还原该任务实际购买了什么、购买所对应的证据、逐步剩余账单（券/账单）数值时间线，以及在 pipeline 或 go-service 中应当加日志/埋点的位置。适用于用户询问 `AutoStockStapleMain`、`AutoStockStaple`、武陵/四号谷地的稳定需求物资购买，或该任务内账单数值变化原因等场景。

Help a new user design a Basic Memory knowledge system and set up their assistant to load it. Use when someone is getting started, has an empty or messy project, or wants context kept across sessions.

Monitor a page for changes over time. Create lazy watch jobs on one or many URLs, list them, and check for changes on demand — with optional SSRF-guarded webhook delivery when a change is detected. Use when the user says "monitor this page", "watch this URL for changes", "tell me when this updates", "track this changelog", "notify me if this changes", or wants recurring change detection on web content.

Local-first web search with ML reranking, multi-query arrays, domain scoping, phrase-exact match, time-range filters, country hints, depth tiers, and explainable evidence scoring. Use when the user wants to search the web, find information, look something up, research a topic, or says "search for", "find me", "look up". Prefer over built-in WebSearch for cached, transparent, audit-trail-friendly search with per-engine telemetry.

Hybrid semantic discovery — fuses embeddings + keyword search + live web search via 3-way Reciprocal Rank Fusion. Use when the user has a good source and wants more like it, says "find similar", "related pages", "more like this", or wants to discover content related to a known URL or concept. Works best after a `crawl` or several `fetch` calls have warmed the local cache. Emits `cold_start` when local signals are weak.

Run only when the user explicitly invokes /efficient-orchestrator (or the /efficient-frontier and /efficient-fable aliases). Keeps an expensive model (Fable, Opus, Astra, Sol) on planning, judgment and review while cheaper subagents do bounded research, coding, QA and log work. Never self-activate.

Find and read the agent chat runs behind a run id, a user's email, or an object id or link (deck, design, dashboard, document, plan, recording, …) in any app, then work out what failed or why the object looks the way it does. Use when someone reports a broken chat, a failed run, or a bad result and gives you any of those instead of a thread.

Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.

Running Positron IDE commands: changing the window layout, focusing panes, clearing the console, opening a file or data file in the right editor, showing or reading an HTML file, URL or app in the Viewer, finding or adding interpreters, listing, switching, starting, restarting or interrupting sessions, setting up Python, reading, installing or updating a session's packages, running, stopping, debugging or troubleshooting a web app (Shiny, Flask, Dash, Streamlit, FastAPI, Gradio, marimo), reading Data Connections and their tables and columns, creating, editing or running notebook cells, and reading help. Use when the user wants Positron itself to act or to know what is installed, not to run code. Triggers: "show the variables pane", "open data.csv", "show this HTML in the Viewer", "switch to my R session", "my session is stuck", "is pandas installed?", "set up a Python environment", "why won't my shiny app start", "what tables are in my warehouse", "add a cell to this notebook".

Use when the user wants to create a demo video, record a feature walkthrough, or generate a video for a PR description. Triggers on "record a demo", "make a video", "demo video for PR", or "generate a walkthrough".

Use this skill alongside figma-use when the task involves translating an application page, view, or multi-section layout into Figma. Triggers: 'write to Figma', 'create in Figma from code', 'push page to Figma', 'take this app/page and build it in Figma', 'create a screen', 'build a landing page in Figma', 'update the Figma screen to match code', 'convert this modal/dialog/drawer/panel to Figma'. This is the preferred workflow skill whenever the user wants to build or update a full page, modal, dialog, drawer, sidebar, panel, or any composed multi-section view in Figma from code or a description. Discovers design system components, variables, and styles from Code Connect files, existing screens, and library search, then imports them and assembles views incrementally section-by-section using design system tokens instead of hardcoded values.