agentscope-ai

Use when admin asks to start a multi-worker project, when a Worker @mentions you with task completion in a project room, when project plan changes are needed, or when a blocked task needs resolution.

Use when admin requests hand-creating or resetting a Worker, starting/stopping a Worker, managing Worker skills, enabling peer mentions, or opening a QwenPaw console. Use agentteams-find-worker only as a helper for Nacos-backed market import or when task…

Execute AI coding CLI tools (Claude Code / Gemini CLI / qodercli / Qwen Code / OpenCode) on behalf of Workers. Use when a Worker sends a coding-request: message, asking the delegating agent (Manager, team Leader, or an authorized Worker) to run coding…

Discover and install agent skills from the open ecosystem. Use when you encounter an unfamiliar domain, framework, or workflow that you lack specialized knowledge about, or when your coordinator suggests searching for skills before starting a task.

Discover and call MCP Server tools via the mcporter CLI. Use when your coordinator notifies you about new MCP tools, or when you need to call external APIs. Includes workflow for generating skill documentation for new MCP servers.

Use for TeamHarness message delivery details after a flow has decided a message must be sent: current-session reply, NO_REPLY, cross-room/cross-channel message tool payloads, PROJECT_REQUESTED task-room handoff, and requester replyRoute reports. Do not use to…

Use for TeamHarness shared workspace paths, filesync, artifact publication, and Matrix room file boundaries, including artifact publish_file and task/project deliverables. Do not use for message routing or project state changes.

Use to understand TeamHarness team roles, responsibility boundaries, and who owns project, task, runtime, and control-plane work.

Create a focused workspace Skill from reusable decisions, knowledge, templates, or workflows in the current conversation. Use for /make-skill with a focus argument and requests such as save this workflow or turn this into a skill; do not use for one-off…

将当前对话中可复用的决策、知识、模板或工作流创建为聚焦的 workspace Skill。适用于带 focus 参数的 /make-skill，以及“保存这个流程”“把它做成 skill”等请求；不适用于一次性总结或普通文件创建。

为 Creator 编译制作级视觉资产图、电影分镜图和参考生视频 Prompt。适用于 角色身份板/设定图、关键帧或多宫格分镜、短片动作节拍、R2V/I2V/T2V 提示词， 尤其适合要求角色一致性、镜头语言、参考图职责、Seedance 时间段或声音设计的 任务。不用于纯素材剪辑、只查询模型参数，或只评价已经生成的成片。

为角色/场景/道具设计视觉资产、编写或修复 VisualVariant 与阵容图的生成 Prompt 时使用。涵盖参考图布局（艺术身份板/序列关键帧/环境研究板/道具研究 板）、构图与镜头语言、图片审核误判规避、实体 continuity 与以主图为参考 的跨 Variant 衍生设计。不用于视频 Prompt 编译（见 professional-media-prompts）、 纯素材剪辑或只评价已生成产物。

数学题分步讲解视频的领域知识。适用场景：(1) 用户给出一道数学题并要求做讲解/ 解题视频；(2) 用户说"讲解这道题"、"生成解题视频"、"make a math tutorial"； (3) 需要中文数学教学视频，涉及方程、公式或几何图形。本 skill 只提供领域 知识（题目拆解、讲解节奏、旁白文案规范、版式与 KaTeX 排版风格）；画面、 配音与成片一律用 Creator 原生 Element 与工具从源头构建。

为“非正式发布会”官方模板设计或修改全屏视频内的中英双语花字，包括渐变材质字、 流光、关键词动效、造型标签与实片排版。仅在选用该模板或用户明确要求该模板的 字幕风格时使用；不改变普通字幕的默认样式，不负责生成底层视频或配乐。

Use this skill whenever you need the allowed_tools / skills preset for a sub-agent role before calling spawn_subagent. Trigger especially when user mentions: "role", "spawn_subagent", "allowed_tools", or any OMP role name (executor, architect, analyst,…

Use this skill only for scheduled or recurring tasks. Manage jobs with qwenpaw cron list/create/get/state/update/pause/resume/delete/run, and always pass --agent-id explicitly.

提供数据分析任务执行的强制性通用规范。在计划完成准备执行数据分析任务、或执行过程中涉及产物落盘、数据复用、异常处理、计划调整、质量自检时，必须优先调用此技能。

将 BI 数据分析结果组织成可视化 HTML 报告。当分析完成、需要生成报告时调用。

将用户的数据分析需求转化为结构化的分析计划并与用户确认。用户提出 BI 业务分析、数据探索、统计建模等需要规划的分析任务时使用。简单数据查询、公式明确的定量计算等流程简单、无需规划的任务不使用。

语义层工具使用规范。在发现语义层工具可用、需要查询指标定义/维度信息、验证业务口径、处理指标歧义时，必须优先调用此技能。无语义层工具时不调用。

通过量化历史数据的自然波动幅度，自适应计算判定阈值。当需要从数据本身确定阈值（如波动阈值、影响度阈值等）、而非使用固定值时调用。仅适用于日/周粒度阈值确定。

基于阈值检测时间序列中的显著异常波动点。当需要找出指标异常波动日期、识别数据异动时调用。

计算各维度（组）值对指标变动的贡献度，支持可加型量值指标和加权平均型/率值指标。当需要计算贡献度、解释指标"为什么涨/跌"时调用。

从运营周报、活动文档、对话输入或文档工具 API 中提取业务事件，与指标异常时间窗口对齐，生成有证据支撑的因果归因假设并排序。当已知指标存在异常波动、需要从外部文档证据中解释"为什么"时调用。

Expert Java developer skill for AgentScope Java framework - a reactive, message-driven multi-agent system built on Project Reactor. Use when working with reactive programming, LLM integration, agent orchestration, multi-agent systems, or when the user…

Four-layer skill composition, skill marketplaces, the self-learning loop

四层技能合成、技能市场、自学习闭环

Statistical data analysis skill — use when the user asks to analyze numbers, compute statistics, or summarize datasets.

Starter skill template. Replace the body with your own playbook before relying on it; this stub exists only to teach the workspace layout.

First-response playbook for an inbound customer message — classifies the ticket, restates the issue, gathers missing context, and routes the reply.

Produces a citation-grounded summary across one or more workspace source files for summaries, briefings, literature reviews, or comparisons.

Use when the user asks for a summary, TL;DR, or condensed version of any content.

Use when the user wants help with academic papers or citations but it's unclear which specific workflow fits — reviewing a paper, checking a BibTeX file for fake references, or benchmarking multiple LLMs on reference-recommendation accuracy. Also use when the…

Review academic papers for correctness, quality, and novelty using OpenJudge's multi-stage pipeline. Supports PDF files and LaTeX source packages (.tar.gz/.zip). Covers 10 disciplines: cs, medicine, physics, chemistry, biology, economics, psychology,…

Verify a BibTeX file for hallucinated or fabricated references by cross-checking every entry against CrossRef, arXiv, and DBLP. Reports each reference as verified, suspect, or not found, with field-level mismatch details (title, authors, year, DOI). Use when…

Benchmark LLM reference recommendation capabilities by verifying every cited paper against Crossref, PubMed, arXiv, and DBLP. Measures hallucination rate, per-field accuracy (title/author/year/DOI), discipline breakdown, and year constraint compliance.…

Use when the user wants to compare or benchmark multiple LLMs/agents arena-style but it's unclear which specific workflow fits — a general-purpose win-rate comparison on a custom task, or a benchmark specifically about reference/citation hallucination rate.…

Automatically evaluate and compare multiple AI models or agents without pre-existing test data. Generates test queries from a task description, collects responses from all target endpoints, auto-generates evaluation rubrics, runs pairwise comparisons via a…

Benchmark LLM reference recommendation capabilities by verifying every cited paper against Crossref, PubMed, arXiv, and DBLP. Measures hallucination rate, per-field accuracy (title/author/year/DOI), discipline breakdown, and year constraint compliance.…

Build custom LLM evaluation pipelines using the OpenJudge framework. Covers selecting and configuring graders (LLM-based, function-based, agentic), running batch evaluations with GradingRunner, combining scores with aggregators, applying evaluation strategies…

Use this skill when the user explicitly wants to connect to a running Chrome browser, scan local CDP ports, specify a `cdp_port`, or share a single browser across multiple agents/tools. By default, browser_use already launches the browser using managed CDP;…

当用户明确希望连接到已运行的 Chrome 浏览器、扫描本地 CDP 端口、显式指定 `cdp_port`，或让多个 agent / 工具共享同一个浏览器时，使用本 skill。当前 browser_use 默认已使用 managed CDP 启动浏览器；如果用户不希望暴露浏览器历史、Cookies 等敏感信息，推荐改用 `private_mode=true` 的隐私模式。

Use this skill when the user needs to control the browser launch mode for browser_use. By default, browser_use launches the local Chrome/Chromium using managed CDP; `headed` controls whether the window is visible, and `private_mode` controls whether CDP is…

当用户需要控制 browser_use 的浏览器启动方式时，使用本 skill。当前 browser_use 默认使用 managed CDP 启动本地 Chrome/Chromium；`headed` 控制是否显示窗口，`private_mode` 控制是否禁用 CDP、改走 Playwright，`browser_args` 传入额外的 Chromium 启动参数，`executable_path` 指定自定义浏览器可执行文件路径。

Use this skill when you need to consult another agent, ask for help, or involve a specific agent the user asked for.

当你需要咨询其他 agent、寻求帮助，或用户明确要求某个 agent 参与时，使用本 skill。

Use this skill only for scheduled or recurring tasks. Manage jobs with qwenpaw cron list/create/get/state/pause/resume/delete/run, and always pass --agent-id explicitly.

仅在需要未来定时执行或周期执行任务时，使用本 skill。使用 qwenpaw cron list/create/get/state/pause/resume/delete/run 管理任务，并始终显式传入 --agent-id。

AI驱动的Java代码安全审计技能，实现系统化、高覆盖率的漏洞挖掘。使用场景： (1) 审计Java/Kotlin项目寻找安全漏洞（0day挖掘、代码审计、安全评估） (2) 企业级代码库的安全审计（支持大型项目） (3) 需要高质量、低幻觉率的安全审计报告 (4) CI/CD集成的前期漏洞发现 触发关键词：Java审计、代码审计、安全审计、漏洞挖掘、0day、安全评估、Java security audit、code review for security

Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling…

Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting,…

Use this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files). Triggers include: any mention of "Word doc", "word document", ".docx", or requests to produce professional documents with formatting like tables of…

Use this skill any time a .pptx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx file (even if the extracted content will be used…

生成短视频脚本和分镜。输入产品/主题描述、目标受众、时长，输出完整脚本+分镜结构（JSON/YAML）。支持多种模板：痛点-解决、反转剧情、before-after。触发词：视频脚本、短视频文案、分镜、storyboard、video script。

A library for building, validating, visualizing, and serializing dialogue graphs. Use this when parsing scripts or creating branching narrative structures.

This skill should be considered when you need to answer reflow machine maintenance questions or provide detailed guidance based on thermocouple data, MES data or defect data and reflow technical handbooks. This skill covers how to obtain important concepts,…

Recall past conversations, preferences, project history, and decisions from ReMe in Qoder, or check the ReMe connection.

Recall past conversations, preferences, project history, and decisions from ReMe in Claude Code, or check the ReMe connection.

Recall past facts, preferences, and decisions from ReMe; inspect memory delivery, service health, and the consolidation schedule; or consolidate memory when requested.

Set up and use ReMe as a file-native long-term memory system through the reme CLI. Use when an Agent needs to detect whether ReMe is installed or running, install and configure ReMe, start or verify its local service, retrieve prior context, or write and…

面向中文自然语言的 Tushare 数据研究技能。用于把“看看这只股票最近怎么样”“帮我查财报趋势”“最近哪个板块最强”“北向资金在买什么”“给我导出一份行情数据”这类请求，转成可执行的数据获取、清洗、对比、筛选、导出与简要分析流程。适用于 A 股、指数、ETF/基金、财务、估值、资金流、公告新闻、板块概念与宏观数据等研究场景。

钉钉消息发送技能。支持企业内部机器人（批量单聊/群聊）和 Webhook 自定义机器人两种接入方式，支持多机器人管理，支持文本、Markdown、链接、ActionCard、FeedCard等多种消息类型。

Search the public web through Serper instead of the built-in WebSearch tool. Use when an agent needs current information, general web sources, or URLs not available from local files and project APIs. SERPER_API_KEY is available in the environment.

qwenpaw_memory

Guidelines for handling CSV/Excel files

Guidelines for handling image files

Guildlines for handling json files

Guidelines for handling databases

Guidelines for handling text files

Participate in DojoZero prediction games, including NBA and FIFA World Cup trials. Use when user wants to find games, join them, check scores/odds, place bets or prediction-window picks, or view leaderboards.

Participate in DojoZero prediction contests. Use when user wants to find games, join prediction-mode trials, check event status, submit predictions, or view leaderboards.

Generates English pull request titles and bodies by comparing the current branch with main. Use when the user asks to draft PR content, summarize branch changes, or produce a PR title/body from git diff.

Run HarnessJevBench fresh inference through HTTP services or external model-owned Python engines, reproduce the published leaderboard, generate label-free inputs, or score supplied predictions. Use for HarnessJevBench evaluation requests in this repository.