Explore

Create, use and clean up cua sandboxes (disposable Linux or macOS computers) locally or in the Cua cloud with the `cua` CLI or the cua SDK, and browse the web inside one. Use when a task needs an isolated machine to run code, test an app, drive a desktop GUI, browse or test a website, fill a web form, take a web page screenshot, or reproduce something without touching the user's own computer.

Drive a native GUI app (macOS, Windows, Linux) via the cua-driver CLI (default) or MCP server; snapshot its accessibility tree, act through snapshot-bound element tokens, native menu paths, exact window geometry, or pixel coordinates, and verify from fresh state. Use when the user asks you to operate, drive, automate, or perform a GUI task in a real application on the host, or to continue, resume, or recall recent Cua activity.

Poll and rank open GitHub issues, RFCs, and pull requests for maintainer work, or start one explicitly selected item. Use when a user asks what to work on, requests backlog priorities, says "poll work," wants actionable issues or pull requests, or says "start

Triage the GitHub privately-reported vulnerability backlog for Activepieces — pull repository security advisories from the Security tab, scope-check against SECURITY.md, deeply validate each, compute SLA status, and propose fix plans for review. Use when the user asks to triage security advisories, work the reported-vulnerability backlog, or check SLA on reported vulns. For Dependabot dependency alerts, use the triage-dependabot-alerts skill instead.

Scan an Activepieces Docker image with grype for OS/base-image (deb) and application (npm) CVEs of High/Critical severity. Lists the 3 most-recent published tags and lets the user pick which to scan, validates each finding is real and reachable, and proves candidate fixes in an isolated git worktree (rebuild image + re-scan + tests + container smoke-run + codebase diff) before proposing anything — fixes are NEVER auto-applied; the user decides per finding. Use when the user asks to grype-scan the image, triage container/image vulnerabilities, or check a shipped Docker image for CVEs.

Triage Dependabot dependency vulnerability alerts for the Activepieces repo — pull open alerts, dedupe to distinct (package, advisory), confirm the vulnerable package + API is actually used, and propose version-bump fixes proven non-breaking (build + tests) before any PR. Use when the user asks to triage Dependabot alerts, work the dependency-vulnerability backlog, or bump vulnerable dependencies. For human-reported vulnerabilities, use the triage-security-advisories skill instead.

Review code for logging patterns and suggest evlog adoption. Guides setup on Nuxt, Next.js, SvelteKit, Nitro, TanStack Start, React Router, NestJS, Express, Hono, Fastify, Elysia, oRPC, Cloudflare Workers, and standalone TypeScript. Detects console.log spam, unstructured errors, and missing context. Covers wide events, structured errors, drain adapters (Axiom, OTLP, HyperDX, PostHog, Sentry, Better Stack, Datadog), sampling, enrichers, and AI SDK integration (token usage, tool calls, streaming metrics, telemetry integration, cost estimation, embedding metadata).

Diagnose a worker that is growing memory or being OOM-killed: separate a JS-heap leak from native growth, take a V8 heap snapshot of a live worker in place (never uploading it), and walk the retainer path back to the code that holds the memory. Use when a self-hoster reports 'memory correlates with pod age', workers restart on their own, or a container sits near its limit.

Find what is burning CPU in a live app container without restarting it: separate the JS thread from GC and libuv, open the V8 inspector in place with SIGUSR1, take a CPU profile over CDP, and aggregate it back to the function that owns the time. Use when containers sit at their CPU cap, healthchecks time out with nothing crashed, or a self-hoster reports the app 'unhealthy' but the process is alive.

Generate `outputSchema` for an Activepieces piece's actions and triggers, so a step's output renders as a curated, labelled tree in the flow builder and data selector. Use when the user asks to add or improve outputSchema for a piece.

Build and edit Activepieces pieces (integrations) — creating new pieces, adding actions or triggers, or fixing bugs in existing ones. Use when the user asks to work on an Activepieces piece, connector, or integration.

Build and maintain documentation sites with Mintlify. Use when creating docs pages, configuring navigation, adding components, or setting up API references.

Design system for Activepieces (open-source AI automation platform, "open source replacement for Zapier"). Use whenever designing, mocking, or building UI for Activepieces — the web app (flow builder, runs, connections, dashboard), docs, or marketing surfaces. Provides brand purple `#8142E3`, Inter type ramp with `sm` (14px) as body default, Tailwind neutrals, Lucide icons, Shadcn/Radix primitive conventions, the signature dotted-canvas builder background, and a recreated web UI kit.

Debug a failed Activepieces flow run end-to-end: given a flow run id (or BullMQ job id), find why it failed, cross-referencing the live BullMQ job + Postgres rows (SSH script on the DevOps box), the centralized ClickHouse logs (ClickStack MCP), and the code in this repo, then categorize the failed-job backlog on request.

Build or review audit trails in TypeScript/JavaScript apps using evlog (pipelines, typed actions, denials, retention, compliance-style reviews). For application code, not for extending the evlog package.

Analyze application logs from the .evlog/logs/ directory. Use when debugging errors, investigating slow requests, understanding request patterns, or answering questions about application behavior. Reads structured NDJSON wide events written by evlog's file system drain.

Control a real browser via CDP: clicking, typing, navigation, logged-in sessions, JS-rendered or bot-protected pages. Not for plain HTTP fetches of public content - use curl for those.

Generate and edit images, vector artwork, photos, audio and video; inspect results and deliver finished artifacts.

Create, inspect and deliver charts, PDFs, Word documents and spreadsheets, and open animated browser presentations.

Investigate, implement and verify code changes directly in a repository using Leon's file, search and shell tools.

Writing and maintaining lat.md documentation files — structured markdown that describes a project's architecture, design decisions, and test specs. Use when creating, editing, or reviewing files in the lat.md/ directory.

Expert in building cross-platform desktop applications using web technologies (HTML/CSS/JS) with the Electron framework.

Turn documents, notes, transcripts, and meeting records into traceable action items, decisions, owners, deadlines, and follow-ups without inventing missing commitments.

可靠工程工作流的执行细则——探索→计划→TDD→系统化调试→代码审查→完成前验证。供方法论教练在开发任务中逐步引导。

使用 akshare-cli 获取中国及全球金融市场数据（1090+ 个函数）。覆盖股票、基金、期货、债券、外汇、宏观经济、指数、期权、新闻等全领域。当用户需要以下任何一种数据时触发此 skill：股票行情、历史K线、实时报价、涨停跌停池、龙虎榜、板块资金流向、基金/ETF净值、期货行情、债券/可转债、外汇汇率、宏观经济指标（GDP/CPI/PMI）、指数数据、期权数据、财经新闻。关键词触发：akshare、股票、行情、K线、基金、期货、外汇、宏观、国债、可转债、涨停、龙虎榜、板块、资金流向、stock data、market data、financial data、A股、港股、美股。即使用户只是随口提到"看看某个股票"或"最近市场怎么样"，也应触发。

为用户从指定新闻网站查找最新新闻。提供政治、财经、社会、国际、科技、体育和娱乐类别的权威 URL。使用 browser_use 打开每个 URL 并通过 snapshot 获取内容，然后为用户总结。

帮助用户了解 Octop、解答产品与使用问题，并配置当前实例。当用户提出以下类型的问题时使用此 skill： Octop 是什么、能做什么、有哪些功能或亮点；官网、帮助文档、安装、快速开始、Docker； 专家、人格、知识库、连接器、通道、定时任务、ACP、远程桌面等概念或求助； 配置或切换 LLM 模型与 Provider；添加或管理 IM 通道（飞书、企业微信、QQ 等）； 启用或禁用 Agent Skill；管理定时任务；备份与升级；询问「octop 怎么用」「怎么配置」、 「怎么接入 xxx」「怎么换模型」「怎么加通道」「CLI 怎么用」「有没有文档」等。 即使用户只是问「Octop 是什么」或「怎么配置 octop」，也应触发此 skill。

学习订阅任务创建与启用。医生完成 5 项登记后，按用户选择创建订阅任务（每日指南连续学习/专业指南更新提醒/地区医保政策变化学习）。平台 cronjob_create 自动绑定当前会话通道投递，无需微信绑定门禁。先查重再创建，回执如实写已启用。

医学信源快速或完整核验。普通教育性医学问答用单一权威正文快速核验；最新版本、精确推荐、高风险药品、医保监管或跨版本请求升级为完整核验。只整理学习依据，不生成诊疗、处方、急诊处置、医保报销结论或HIS执行建议。

结构化医学学习产物的统一格式校验。四条硬格式与 validate_output.py 单次标准输入流程。指南章节、路径图、诊断、备考、医保、更新和每日单元使用；普通医学问答走 clinical-q-and-a 的轻量自检，通用任务不套用。

Route retrieval of Chinese medical guidelines, consensus statements, regulatory documents, drug information, safety notices, standards, and public-health materials through authoritative originals and verified fallback copies. Use when a canonical medical source is blocked, slow, unavailable, moved, paywalled, or difficult to parse, or when designing and auditing domestic medical-source fallback rules. Do not treat secondary summaries or reposts as independent clinical evidence.

请求入口的意图识别与技能路由。按渐进式判断意图：先通用任务（直接做）、再高风险临床（拒绝）、再区分纯信源核验与医学学习、最后登记/订阅。纯信源核验先走 source-verify；原始路径受限时再进入 medical-source-failover。医学学习才调对应 skill + output-format 校验。

为基层医生学习小助手生成地区医保政策学习内容。支持每周二、周五或按需的最新变化摘要，也支持近一年／指定区间的重点政策回顾学习。只输出政策原文摘要、待本地确认事项和权威信源，不输出报销结论、支付比例、HIS提醒、病历要求或医院执行承诺。

为基层医生学习小助手生成专业指南更新提醒。适用于每3天一次或按需输出与医生科室、亚专业、职称深度和基层场景相关的指南、共识、规范、临床路径或质控标准更新。只提供学习提醒和权威信源，不输出临床行动、疾病SOP、个体患者建议、药物剂量或院内流程变更。

把已核验的权威指南按原文结构逐节展开讲解，并给出原文定位，帮医生读懂原文；也可把指南整理为学习路径图（学习顺序）。不得生成诊疗方案、床旁处置步骤、疾病 SOP、急诊行动卡、个体患者建议或药物剂量说明。

当用户要求评估自己对某份指南的掌握情况、制定指南学习/备考计划、梳理薄弱章节、 做教育性小测，或学习指南中的诊断标准、鉴别框架和流程关系时使用。 本技能只做学习诊断与诊断标准教学，不对具体患者作疾病诊断、治疗、处方、剂量、 急诊处置或医保报销判断。

指南学习流程约束。判断请求类型→读取状态→按流程创建轨道/预览/投递/排查迁移；操作命令与细则见 references/learning-operations.md。选题用权威信源，输出前经 output-format 校验。不生成诊疗/处方/急诊SOP/个体患者建议。

根据医生自述的考试或岗位目标（副高、中级、规培、执业医师等）推荐学习材料、教材与学习顺序，并按目标动态调整学习内容。必须以官方考试大纲为准，不得断言考试范围或承诺通过结果，也不得输出个体诊疗内容。

当医生提供或更新称谓、地区、医院、科室、职称这5项登记信息，变更科室/地区/职称，或要求停用、删除订阅数据时使用。本技能只管登记的写入/更新/清除；登记完成后的订阅任务创建由 subscription-setup skill 负责。不启用诊断、治疗、疾病SOP、急诊行动卡、个体患者建议、药物剂量、HIS决策支持或医保报销结论能力。

快速回答普通教育性医学问题与常见误区。默认用一份国内现行权威指南/共识完成轻量核验并立即作答；不用于个体诊疗、处方剂量、精确推荐定位、药品高风险事实、医保监管或跨版本比较。

AI 工具安全部署与合规体检的 6 步清单，覆盖 shellward 8 层防御与中文合规（PIPL/等保2.0/数据出境/生成式AI标识）。

AI 编程工具速查与 60+ 技巧清单——覆盖 Claude Code / Cursor / Codex / Copilot / Aider 等的能力矩阵、核心命令与高频技巧分类。供编程导师在选型与排障时引用。

Use a concise, polite greeting and opening. Trigger this skill when the user says hello, asks for a self-introduction, or wants a more formal / polite tone.

Publish the Octop Python package: cut a release branch from develop, bump version, update CHANGELOG, open a PR to main; after merge, Actions tag on main (PyPI / Docker Hub + GHCR) and sync main into develop. Use when the user asks to publish, release, bump version, cut a release, or run /publish.

Compare multiple sources on a topic and produce a grounded comparison matrix. Use when the user asks to compare papers, tools, approaches, frameworks, or claims across multiple sources.

Search prior Feynman session transcripts on disk. Use when the user asks about earlier research sessions, past findings, or wants to resume a prior session.

Write a durable session log capturing completed work, findings, open questions, and next steps. Use when the user asks to log progress, save session notes, write up what was done, or create a research diary entry.

Run a tough but constructive internal research critique of an AI research artifact. Use when the user asks for a review, critique, feedback on a paper or draft, or wants to identify weaknesses before submission.

Plan a replication of a paper, claim, or benchmark, and execute only after an explicit environment choice. Use when the user asks to replicate results, reproduce an experiment, verify a claim empirically, or build a replication package.

Read, extract, and cross-check content across scientific PDFs. Use when a task needs methods, figures, tables, citations, accessions, or claims from multiple places in one or more papers.

Turn research findings into a polished paper-style draft with sections, equations, and citations. Use when the user asks to write a paper, draft a report, write up findings, or produce a technical document from collected research.

Compare a paper's claims against its public codebase. Use when the user asks to audit a paper, check code-claim consistency, verify reproducibility of a specific paper, or find mismatches between a paper and its implementation.

Find implementable ML training recipes from papers, datasets, docs, and code. Use when the user wants to fine-tune, train, reproduce, or choose a practical ML method, dataset, hyperparameter setup, or benchmark recipe.

Search, read, and query research papers via Feynman's alphaXiv-backed alpha tools. Use when the user asks about academic papers, wants to find research on a topic, needs to read a specific paper, ask questions about a paper, inspect a paper's code repository, or manage paper annotations.

维护 AutoClip 的公开社区看板。当用户要看社区反馈、整理功能呼声、把一条讨论放进路线图、 或问 GitHub Discussions、Projects、Quackback、Featurebase 该怎么接时使用。 想法留在 Discussions，只有人确认后才开 Issue。

用 AutoClip 把一条 YouTube / B 站链接，或本地长视频（讲座、访谈、播客、课程），做成各平台成片、封面、发布文案和 ZIP 发布包。 当用户要「出片」「做封面」「写发布文案」「打包发布」「从这个链接做短视频」时使用。 只要原始高光切片、合集，或把旧切片发到 Upload-Post 时不要用这份说明，改读 references/legacy.md。

Retrieve and analyze World Bank indicators, country metadata, and development time series.

Build or modify production-oriented web applications by following the repository architecture and validating real user flows.

Create, validate, package, and optionally install original KunAgent skills.

Retrieve and analyze public SEC filings, company facts, XBRL data, and filing metadata.