Selects, deploys, and customizes AI models on Amazon SageMaker. Training or Processing jobs, fine-tuning (SFT/DPO/RLVR/RLAIF), model selection, dataset preparation, evaluation, SageMaker or Bedrock deployment, inference optimization and endpoint diagnostics. Covers the full lifecycle from planning through production. Use when fine-tuning models on SageMaker, choosing which base model to customize, fine-tune, or deploy from SageMaker JumpStart or Hub, SageMakerPublicHub, or the SageMaker public model catalog, transforming or validating training data, evaluating model quality, deploying or optimizing endpoints, configuring IAM/S3 for training, or managing SageMaker Managed MLflow. Use for endpoint health, failures, latency, logs, metrics, errors. Covers Serverless Model Customization, Nova and OSS deployment paths, and PySDK v3. NOT for Ground Truth labeling, Feature Store, or general-purpose AWS infrastructure.

Stores and retrieves genomic variant calls with TileDB-VCF. Use for indexed single-sample VCF/BCF ingestion, incremental cohorts, region and sample queries, streaming results, allele statistics, QC, and VCF/BCF export locally or through TileDB Cloud.

Creates product search agents with semantic search and RAG on Google Cloud (Vector Search on Gemini Enterprise Agent Platform, BigQuery, embeddings). Use when the user wants to "build a product search agent", "create an e-commerce search", "make a shopping assistant", "set up semantic catalog discovery", "ingest products into Vector Search", or "deploy a retail RAG agent". Handles the full pipeline: catalog data ingestion to BigQuery, Gemini Enterprise Agent Platform Vector Search collection setup, ADK agent scaffolding, evaluation, and Cloud Run deployment.

Integrate Mem0 into an existing repository using a goal-driven, TDD pipeline. Detects the repo's language automatically and asks the user to pick between Mem0 Platform (managed) and Mem0 Open Source (self-hosted). Writes failing tests before any implementation. Produces a local feature branch plus `.mem0-integration/` artifacts consumed by the paired verification skill. TRIGGER when: user says "integrate mem0", "add mem0 to this repo", "wire mem0 into <repo>", or asks how to add memory to an existing project. DO NOT TRIGGER when: the user wants general SDK usage (use skill:mem0), CLI usage (use skill:mem0-cli), or Vercel AI SDK (use skill:mem0-vercel-ai-sdk). After success, invoke skill:mem0-test-integration to verify in the same workspace (loose coupling).

Generate and A/B test Google Ads copy. Use when asked to write ad copy, headlines, descriptions, create ad variants, test ad messaging, improve CTR, or generate RSA (Responsive Search Ad) components. Trigger on "ad copy", "write ads", "headlines", "descriptions", "RSA", "responsive search ad", "ad text", "ad creative", "improve CTR", "ad A/B test", "ad variants", "write me an ad", "ad variation experiment", or when the user wants to improve click-through rate on existing ads.

Plan, validate, and safely publish Google Ads assets, including sitelinks, callouts, structured snippets, image assets, and Performance Max asset briefs. Use when asked for Google Ads assets, ad extensions, sitelinks, callouts, snippets, image assets, Performance Max assets, PMax creative, or an asset audit.

为 CC GUI 桌面客户端创建、修改、迭代插件（Tier-0 声明式 CSS/JSON，或 Tier-1 单文件 ESM）。当用户要"做一个插件 / 给 CC GUI 加个功能 / 改输入框或界面样式 / 加设置页、侧边栏或状态栏入口 / 加面板页签或命令面板命令 / 把某套工作流包装成插件 / 改一下我做过的那个插件"，或提到 ccgui 插件、插件目录、从本地目录安装插件时使用。

Implement a VS Code Sweeper fix spec — fetch the sweeper's agent-fixable review record from its public state repo, implement the narrow fix in the current vscode checkout, and — after showing the diff — open a draft PR. Use ONLY when the request explicitly asks for the sweeper — "sweeper-fix", "sweeper", "vscodesweeper", a sweeper record, or a sweeper fix spec. Do NOT use for a plain "fix this issue" request that doesn't mention the sweeper; fix those directly with your normal tools instead.

Use when investigating a specific named Positron e2e (Playwright) test that is failing or flaking -- in CI ("why is <test> failing on main", "this test is flaky in CI", "is this a test bug or a product bug") or on the engineer's machine ("this e2e test just failed locally", "debug this e2e failure"). From CI it surfaces the test's distinct failure modes from history and pulls evidence for one; locally it reads the run's own trace, snapshot, and logs. Either way it reasons to a falsifiable root cause with the engineer and lands on a test fix, a product-bug issue, or an accepted-flake note. Do NOT use for a test you are writing or currently editing (author-e2e-tests); a whole CI run's failures, or a run ID/URL (e2e-failure-analyzer); Vitest or extension-host failures.

Convert .NET tests from xUnit.net v2/v3 to MSTest v4 while preserving VSTest or MTP. Use for replacing xunit packages, Fact/Theory/InlineData/MemberData, assertions, IClassFixture/ICollectionFixture, ITestOutputHelper, TestContext cancellation, traits/Owner, skips, timeouts, and xUnit parallelization. Also use when a "convert xUnit to MSTest" request may already be migrated: inspect and report the no-op. Do not use for xUnit v2-to-v3, MSTest upgrades, NUnit/TUnit conversion, or runner-only VSTest-to-MTP migration.

Enforce Magic Web structured business error logging when adding, changing, migrating, or reviewing `logger.error` calls in `src`, `enterprise/src`, or `customer/src`, including catch blocks, callbacks, ErrorBoundaries, workers, WebSocket/storage flows, and logger adapters. Use to choose `eventKey/errorKind/error/message/context`, preserve historical diagnostics, prevent unsafe payloads, and verify the JS probe plus `/log-report` behavior.

Resolve which slide, page, and (optionally) selected element the user is currently viewing in the open-slide dev server. Consult this whenever the user references "this page", "this slide", "this element", "the slide I'm on", "the current page", or any deictic reference to slide content without naming it. Re-read `node_modules/.open-slide/current.json` at the start of every such turn — the user navigates between turns, so a value you read earlier in the conversation is almost certainly stale.

Designs and audits PCR and RT-qPCR primers with Primer3, explicit thermodynamic conditions, reference-based off-target amplification searches, and traceable sequence coordinates. Use for designing primer pairs, checking existing primers, exon-junction or isoform-specific assays, variant masking, cloning tails, multiplex compatibility, and interpreting Primer-BLAST results. Includes bounded local in-silico PCR and BLAST screening; distinguishes computational candidates from experimentally validated assays.

Guide for adding a new RPC to personhog-replica and personhog-router. Covers eligibility checks, proto definition, code generation for Python and Node.js clients, Rust implementation (storage trait, postgres queries, service handler, router wiring), and index compatibility validation. Use when adding a new gRPC endpoint to personhog, migrating a Django ORM query to personhog, or extending the personhog service API.

Graph-informed mutation testing triage. Parses codebases with Trailmark, runs mutation testing and necessist, then uses survived mutants, unnecessary test statements, and call graph data to identify false positives, missing test coverage, and fuzzing targets. Use when triaging survived mutants, analyzing mutation testing results, identifying test gaps, finding fuzzing targets from weak tests, running mutation frameworks (including circomvent and cairo-mutants), or using necessist.

Track what top crypto futures traders — "smart money", whales, elite accounts — are doing across Binance, Hyperliquid, Bybit and OKX with TraderSpy's live tracking data. The elite leaderboard by smart score, top traders by ROI or PnL over 24h to 30d, open and closed positions filtered by coin, exchange or trader, one trader's profile and closed-trade history, and aggregate market stats. Use this whenever the user asks what whales or top traders are buying, shorting or holding, who the best traders are on an exchange, whether big accounts are long or short a coin, wants to research a trader's record before acting on it, or says "smart money", "whale watch", "what are the pros doing". Not for funding rate / open interest / long-short ratios of the whole market (technical-analysis) and not for AI signals (trading-signals).

Read one crypto futures pair (or a few) properly with TraderSpy — 19 indicators (RSI, MACD, EMA/SMA stacks, Bollinger, ATR, ADX, Stochastic, OBV, VWAP, CCI, MFI, Williams %R, ROC, SuperTrend, Ichimoku, Keltner), pivot points and swing support/resistance, Fibonacci and volume profile, RSI/MACD divergence, squeezes and candle patterns, up to three timeframes and up to three coins in ONE call with a confluence verdict — plus funding rate, open interest and long/short positioning, live price and OHLCV candles. Use this whenever the user asks about a coin's chart, trend, momentum, "is X overbought / oversold", support and resistance, where the 200 EMA is, "analyse BTC", "how does SOL look", "what do the indicators say", "is this a good entry", whether funding is high or open interest is building, or wants price / candle data — even if they only name a coin and ask how it looks. For scanning MANY coins by conditions use market-screener; for whale positions use smart-money.

Use when the user asks to create, prepare, or set up a RevenueCat experiment (A/B test), to prepare its variant offerings or paywalls, or to start, pause, resume, or stop an experiment via the RevenueCat MCP experiment tools

Check whether a RevenueCat user currently has access to a paid feature via entitlements. Use when the user asks to gate a feature behind premium, check if the user has a pro subscription, read customerInfo active entitlements, show or hide a feature based on subscription status, react to entitlement changes, or 'is the user subscribed' on iOS, Android, Kotlin Multiplatform, Flutter, or React Native.

Three isolated reviewers vote PASS or BLOCK on a high-risk change; the majority decides, and every MUST-FIX any of them raises is fixed. Use for changes to auth, payments, data migrations or anything that deletes data, after a third fast-review FAIL, or when someone asks for an independent panel.