Skills

Solves slider captchas by dragging, with audio challenge fallback

Converts books and documents (PDF, EPUB, DOCX, HTML, Markdown, plain text, RTF, MOBI/AZW with Calibre) into structured agent skills, extracting frameworks, mental models, principles, techniques, and anti-patterns. Use when the user wants to study a document through GitHub Copilot CLI, Amp, or Claude Code, apply an author's frameworks while working, or build a reusable knowledge base from a file.

>-

Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Use when you want clean RL abstractions, easy algorithm experimentation, or scalable training with Monarch and TorchTitan.

Use when the user wants to find problems, audit workflow quality, or get a comprehensive health check on their AI workflow.

Color contrast and visual accessibility specialist. Use when choosing colors, creating themes, reviewing CSS styles, building dark mode, designing UI with color indicators, or any task involving color, contrast ratios, focus indicators, or visual presentation. Ensures WCAG AA compliance for all color and visual decisions. Applies to any web framework or vanilla HTML/CSS/JS.

When the user wants to plan, design, or implement an A/B test or experiment, or build a growth experimentation program. Also use when the user mentions "A/B test," "split test," "experiment," "test this change," "variant copy," "multivariate test," "hypothesis," "should I test this," "which version is better," "test two versions," "statistical significance," "how long should I run this test," "growth experiments," "experiment velocity," "experiment backlog," "ICE score," "experimentation program," or "experiment playbook." Use this whenever someone is comparing two approaches and wants to measure which performs better, or when they want to build a systematic experimentation practice. For tracking implementation, see analytics. For page-level conversion optimization, see cro.

Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.

ePub document accessibility specialist. Use when scanning, reviewing, or remediating .epub files for accessibility. Covers EPUB Accessibility 1.1 (WCAG 2.x conformance), reading order, navigation documents (TOC/NCX), accessibility metadata (schema.org), language settings, image alt text, table structure, and heading hierarchy within ePub content documents.

Use when creating or modifying Model Context Protocol (MCP) servers with FastMCP framework - guides through tools, resources, prompts, authentication, Claude Desktop integration, and production deployment with Python and TypeScript examples

Run a lightweight, blame-free postmortem after an incident, failed launch, or missed deadline — for one person. Use when the user says "that didn't go well", "the launch flopped", "we had an outage", "I missed my deadline", or wants to learn from a recent failure.

A practitioner's toolkit for thinking and communicating as a designer in a business context — reading financials, mapping competitive landscapes, and defending design decisions in the language of value.

Reviews a PR or diff with multi-angle finders and adversarial verification, then reports a findings table, a merge/no-merge recommendation, required followups, and offers to create a follow-up PR. Use when the user types /review-pr [PR# | branch | path].

Batch download academic paper full-text (PDF/XML) from a list of DOIs. Handles 25 DOI prefixes across 19 publisher families via three layered routes: (1) publisher TDM APIs requiring institutional subscription (Elsevier ScienceDirect, Wiley Online, Springer Nature), (2) Open Access sources (Crossref, Unpaywall, OpenAlex), and (3) a browser-based fallback for paywalled publishers without TDM access (ACS, RSC, IEEE, AIP, IOP, APS, Annual Reviews, T&F, Chinese journals). Browser fallback offers two routes — Route A drives the user's logged-in Chrome via the OpenClaw `browser` tool with `profile="user"` (best for interactive sessions), Route B uses the standalone `auto-paper-harvester` CLI with its built-in Playwright (best for unattended bulk runs). Use when the user wants to harvest, scrape, fetch, or bulk-download papers from a DOI list, savedrecs export, or Excel; or wants to fill missing full-text PDFs for an existing literature collection. Triggers on phrases like "批量下载文献", "下载全文", "harvest papers", "scrape full text", "TDM API", "下载 Elsevier 全文", "Wiley 批量下载", "下载 PDF".

|

>-

Expert guidance for GRPO/RL fine-tuning with TRL for reasoning and task-specific model training

Bulk-localize subscription and in-app purchase display names across all App Store locales using asc. Use when you want to fill in subscription/IAP names for every language without clicking through App Store Connect manually.

Alignment conversation starting from a 想法/idea. Co-decides with the user whether the idea should be acted on directly in the current session, or fixed into a formal Task for independent dispatch (one-off or recurring). Handles lightweight 'do it now while we talk', heavyweight 'define precisely, run later or on a schedule', and 'just help me think about this' — all on the same skill. Use when the user arrives via the 想法 panel's 'AI 讨论' button (parameter dictionary in the first message), or says 'let's think this through', 'help me plan this', 'I want to explore X', 'I have an idea', '/task-alignment'. Also use proactively when a user jumps into a complex task without defining scope or success criteria — pause, align, and help them pick the right vessel (this session vs. a task).

Classify, gate, and review changes. Use when landing a PR, releasing, or amending rules. Do not use for failure triage; use night-market-debugging-playbook.

Audit experiment integrity before claiming results. Uses cross-model review (external reviewer backend) to check for fake ground truth, score normalization fraud, phantom results, and insufficient scope. Use when user says "审计实验", "check experiment integrity", "audit results", "实验诚实度", or after experiments complete before writing claims.

Use when opening a PR — produces a clean PR description (what / why / how to verify / risks) from a branch diff against base.

Semantic Design System Skill for Google Stitch. Generates agent-friendly DESIGN.md files that enforce premium, anti-generic UI standards — strict typography, calibrated color, asymmetric layouts, perpetual micro-motion, and hardware-accelerated performance.

Use for anything involving the Synalinks neuro-symbolic LM framework (Keras-inspired) — DataModel/Field/Input, JSON operators (+ & | ^ ~), synalinks.ops, LanguageModel/EmbeddingModel and provider prefixes (openai/anthropic/ollama/groq/openrouter/bedrock/...); the Program class and its four building APIs (Functional/Sequential/Subclassing/Mixed), save/load, summary; generation modules (Generator, ChainOfThought, SelfCritique, Identity, PythonSynthesis) and custom Module subclassing; control flow (Decision, Branch, And/Or/Xor, parallel branches, self-consistency, XOR guards); agents (FunctionCallingAgent, RecursiveLanguageModelAgent/RLM, DeepAgent, Tool, MCP, subagents); KnowledgeBase/RAG (DuckDB, EmbedKnowledge/UpdateKnowledge/RetrieveKnowledge, hybrid search); training (compile/fit/evaluate/predict, callbacks, ProgramCheckpoint); rewards & metrics (ExactMatch, CosineSimilarity, LMAsJudge, ProgramAsJudge, F1Score, custom rewards, masking); optimizers (RandomFewShot, OMEGA/DNS); datasets (gsm8k, hotpotqa, arcagi) and visualization. Synalinks is Keras-shaped, so without guidance LMs mix Keras/LangChain/DSPy syntax — this skill constrains usage to idiomatic Synalinks.

AI image editing prompting patterns for Qwen-Image-Edit. Use when editing photos while preserving identity, reframing cropped images, changing clothing or accessories, adjusting poses, applying style transfers, or character transformations. Provides prompt patterns, parameter tuning, and examples.

Electron and Den, desktop plus cloud, two-sandbox e2e, cloud auth, marketplace, org policy, worker proxy, provider sync, desktop handoff. Validate Electron against a Daytona Den server with unified proof.

HeroUI v2 to v3 migration guide for agents. Use when migrating HeroUI v2 apps to v3, upgrading components, or accessing migration documentation. Keywords: HeroUI migration, v2 to v3, migration guide, upgrade HeroUI.

Programmatic screenshot capture on macOS. Find window IDs with Swift CGWindowListCopyWindowInfo, control application windows via AppleScript (zoom, scroll, select), and capture with screencapture. Use when automating screenshots, capturing application windows for documentation, or building multi-shot visual workflows.

GitLab CLI (glab) wrapper for MR status, issues, and repository operations

React composition patterns that scale. Use when refactoring components with

Google Workspace CLI for Gmail, Calendar, Drive, Contacts, Sheets, and Docs.

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

End-to-end workflow for taking MCP work items from backlog to merged PR. Handles git branching, schema-driven planning, implementation, independent review, and PR creation. Composes spec-quality, review-quality, and schema-workflow skills into a single pipeline. Use when a user says "implement this", "work on this item", "fix these bugs", "pick up the next task", "create a PR for this", "go through the backlog", or references specific MCP item IDs for implementation.

>

Detect IBD-style Distribution Days for QQQ/SPY (close down at least 0.2% on higher volume), track 25-session expiration and 5% invalidation, count d5/d15/d25 clusters, classify market risk (NORMAL/CAUTION/HIGH/SEVERE), and emit TQQQ/QQQ exposure recommendations. Use after market close, before TQQQ exposure changes, or as input to FTD/market-state frameworks. Does not execute trades.

IoT and embedded device security testing methodology. Covers hardware reconnaissance (UART, JTAG, SWD, SPI flash, I2C EEPROM, eMMC chip-off), firmware acquisition (vendor portals, OTA capture, flash dump, binwalk extraction), firmware analysis (filesystem mounting, binary triage, hardcoded secrets, default credential discovery), bootloader attacks (U-Boot console, secure-boot bypass, fault injection), runtime attacks on embedded Linux/RTOS (busybox CVEs, MTD writes, /dev/mem), wireless protocol attacks (Zigbee, BLE, Z-Wave, LoRaWAN, Thread/Matter, sub-GHz), MQTT/CoAP/Modbus/BACnet/OPC-UA exploitation, mobile companion app analysis, cloud-IoT API abuse, and side-channel/glitching basics. Use for IoT pentest, smart-home assessment, ICS/OT testing, or embedded vulnerability research.

When the user wants to generate, iterate, or scale ad creative — headlines, descriptions, primary text, or full ad variations — for any paid advertising platform. Also use when the user mentions 'ad copy variations,' 'ad creative,' 'generate headlines,' 'RSA headlines,' 'bulk ad copy,' 'ad iterations,' 'creative testing,' 'ad performance optimization,' 'write me some ads,' 'Facebook ad copy,' 'Google ad headlines,' 'LinkedIn ad text,' or 'I need more ad variations.' Use this whenever someone needs to produce ad copy at scale or iterate on existing ads. For campaign strategy and targeting, see ads. For landing page copy, see copywriting.

Buttons, inputs, pills, badges, calendars, and other interactive components form a visual family — they share the same border-radius, colour logic, shadow scale, border style, and spacing rhythm. Inconsistency between them breaks the sense of a coherent product. Use when building or reviewing a component library, design system, or any set of UI components.

Save current task list for reuse across sessions

Send notifications to Feishu/Lark. Internal utility used by other skills, or manually via /feishu-notify. Use when user says "发飞书", "notify feishu", or other skills need to send status updates.

AI HOT (aihot.virxact.com) 中文 AI 资讯查询 Skill。当用户想知道"今天 AI 圈有什么"、"AI 日报"、"AI HOT"、"AI 资讯"、"AI 热点"、"最近 AI"、"OpenAI/Anthropic/Google 最近发布了什么"、"AI hot today"、"AI news today"、"看一下 AI 行业动态"、"今天有什么大模型发布"、"昨天 AI 圈"、"看下精选条目"、"AI HOT 精选"、"最近一周的 AI 论文"、"AI 模型发布"、"AI 产品发布"、"AI 行业动态"、"AI 技巧与观点" 等任何中文 AI 资讯查询时使用。即使用户只说"AI 圈"、"AI 新闻"、"AI 日报"，或者只是问"今天发生了什么"且上下文是 AI / 大模型 / LLM / 创业领域，也应该触发本 Skill。Skill 会直接 curl 公开 REST API 拉数据并整理成中文 markdown 简报，不需要用户配置任何 API Key 或 MCP server。**不要 undertrigger**——用户问 AI 资讯而你不调本 Skill 就是把过时的训练数据当作今日新闻，对用户有害。

Persistent memory for Claude Code agents with an agent-audit-ritual architecture. User only talks; the agent captures, audits, proposes promotions, and writes. Memory lives in layers — a hot cache (MEMORY.md) held under three size caps, per-session handoffs (context/handoffs/), topical knowledge articles (knowledge/concepts/), and canonical rules (.claude/rules/) — plus multi-project isolation via projects/<name>/ and an experiments/ sandbox. /close-session runs the end-of-session audit ritual. Zero external dependencies.

Auto-review pipeline — reads the full CEO, design, eng, and DX review skills from disk and runs them sequentially with auto-decisions using 6 decision principles. (gstack)

Add image vision to Deus agents. Resizes and processes WhatsApp image attachments, then sends them to Claude as multimodal content blocks.

Google Chat: Send a message to a space.

Organize qualitative research data into an affinity diagram with themes, clusters, and insight statements. Use when synthesizing large amounts of qualitative data from interviews, observations, or surveys.

>

>-