원클릭으로
OnMyAgent
OnMyAgent에는 WeaveQ에서 수집한 skills 119개가 있으며, 저장소 수준 직업 범위와 사이트 내 skill 상세 페이지를 제공합니다.
이 저장소의 skills
物流单据录入与生成方法论。当需要把客户微信文字、语音、手写单图片等碎片信息整理成物流单、发货单、发车单/派车单或运单,校验字段完整性,合并追问缺失信息,实时更新白红黄三联 HTML 效果图,或在确认后按用户选择分别生成 PDF 与/或 Excel 时使用。
OnMyAgent frontend primitive refactor workflow. Use when editing UI components, aligning same-type component sizes, reducing arbitrary Tailwind classes, migrating repeated button/input/badge/icon/status/loading/code/card patterns into shared primitives, or preventing design-token drift while coding.
OnMyAgent UI regression audit for theme consistency, i18n copy, settings pages, and screenshot-based visual checks. Use when scanning UI regressions, verifying design-system token usage, checking mixed Chinese/English copy, or producing page-by-page visual audit reports.
创业庄园知识模块·商业财务。包含商业模型约束、财务仪表盘、现金流、预算。对应《创业可以学》第5-6章。被 manor-judge 调用,不直接面向用户触发。当用户问题涉及"商业模式/财务/现金流/预算/模型太重/盈利/成本"时由 manor-judge 读取本 skill。
创业庄园知识模块·成长转型。包含手艺磨练、实战环境、转型观、三阶段人生。对应《创业可以学》第11-12章 + 创能量 + 创业的盲点。被 manor-judge 调用,不直接面向用户触发。当用户问题涉及"成长/转型/手艺/学习/阶段/倦怠/职业发展/创能量"时由 manor-judge 读取本 skill。
创业庄园知识模块·客户GTM。包含目标人群、客户选择、GTM框架、销售阶段、关键活动。对应《创业可以学》第3-4章。被 manor-judge 调用,不直接面向用户触发。当用户问题涉及"客户/目标人群/GTM/销售阶段/钩子/客户选择/切入点"时由 manor-judge 读取本 skill。
创业庄园知识模块·人执行组织。包含人是结果、系统自运转、环境设计、舞台交付、沟通、猴子管理。对应《创业可以学》第7-10章。被 manor-judge 调用,不直接面向用户触发。当用户问题涉及"团队/招人/执行/管理/沟通/环境/舞台/系统/盯人"时由 manor-judge 读取本 skill。
创业庄园庄主的判断与回答层。负责具体创业问题的识别、回答组织、一语道破、追问收尾。触发场景:(1)用户提出具体创业问题(明确痛点、场景、困惑) (2)用户问体系根概念(OPC/GTM/目标人群/销售阶段等) (3)用户描述经营现状(忙没结果、客户模糊、销售难、模型重、执行靠盯、团队弱、转型难等)。与 manor-host 互斥触发:本 skill 不负责接待,只负责判断和回答。调用5个 knowledge 模块获取知识支持。
运单与货物实收实发核对方法论。当需要对比托运单/运单与实际收货或发货清点,识别有单没货、有货没单、数量品名包装不符,输出差异点、当场证据清单与核对结果记录时使用。收货当场核对优先。
挂靠车队合规台账方法论。当需要汇总挂靠车辆与司机的驾驶证、从业资格、行驶证、营运证、交强险、商业险、年检、违章等信息,建立统一台账,设置 30/15/7 天到期与过期提醒,预警人车证不符、脱保、违章堆积,或生成催补证话术与单车合规查询时使用。非法律意见。
物流开票与应收作业法。当需要对账开票、整理开票资料(客户名称、税号、地址电话、开户行账号、货物或服务名称、金额、对应运单号)、生成开票申请内容、建立运单与开票对应关系、跟踪开票与回款进度、按账期节点催办、处理抬头变更或金额异议时使用。不替代财务系统数据,不替你决定开票抬头与税率,金额与单号无来源不编造。
车队油费稽核方法论。当需要根据加油记录、行驶里程、油卡流水计算单车油耗,与车型正常范围对比,标记少加多开/非定点/套现嫌疑/时空矛盾等异常,并输出优先级稽核报告时使用。结论为稽核线索,禁止编造流水。
物流应收台账与催收方法论。当需要根据客户账期、开票、回款与运单建立或更新应收台账,计算账龄与到期节点,按到期前/到期日/逾期阶段生成提醒与催款话术,校准催收力度,或标注承兑与坏账风险时使用。禁止非法催收与编造金额。
物流理赔案件作业法。当需要处理货损、延误、丢件等异常,整理理赔材料与证据清单,做责任初步判断,起草客户/司机/保险公司沟通话术,建立报案-定损-补件-谈判-结案进度跟踪,或识别易拖延环节并给出加速动作时使用。非法律意见,禁止伪造证据。
网点仓库存台账方法论。当需要登记入库/出库/移库/盘点,更新库存账面,发现账实不符、货位错误、超期滞留、单据与实物不一致等异常,按货物特性给存放注意,倒查错发漏发,或生成简单库存报表与提醒话术时使用。
专线/零担询价与结构报价方法论。当需要根据线路、重量/体积、车型、时效、装卸、回单等生成参考报价,输出最快/最便宜/平衡多档方案,识别含税/回单/装卸/旺季等隐藏条件,按油费路桥折旧司机分摊等构成测算成本区间,检查附加费漏项,保护底价并生成报价与砍价话术,或沉淀本票指导价备忘时使用。无成本库时不编造真实市价。
动态运力池与多因素配载方法论。当需要汇总车辆位置、剩余载重、司机状态、预计空闲时间与方向偏好,维护实时可用运力清单,标记超时未更新车辆,生成司机群回报/催更话术,或把待调度订单与可用运力做成1-3套配载方案(顺路、载重、意愿、时效、空驶)并说明理由与风险时使用。
回单跟踪与费用对账方法论。当需要汇总回单回收状态、统计超期与催收优先级,汇总运费及附加费,生成对账单草稿,标注金额/票数差异及可能原因,或生成催回单/催补资料话术时使用。
冷运作业:预冷检查、温控异常分级、断链证据链、腐损责任初判、多温区注意。触发:冷链、冷藏、冷冻、断链、腐损、预冷、温控报警。
专线干线作业:抛重计费、网点中转时效拆解、配载拼载启发式、外请车约束、回单账期。触发:专线、中转、抛重、配载、外请、回单周期。
物流异常与索赔初判:破损、延误、拒收、压车、放空。输出证据清单、责任初判、各方话术与闭环单。非法律意见。触发:破损、货损、延误、拒收、压车、放空、索赔、扯皮。
物流报价台:专线/整车/零担结构化报价、费用项拆解、底价保护与成交话术。无市价库时不编假成交价。触发:报价、专线价、整车、运费怎么报、底价、给客户报价。
物流应付/应收对账:承运商账单与我方运单应付比对,输出差异表、原因初判、催收或付款清单。触发:对账、账单、应付、回单、票差、该付该催。
短途城配:多点顺序启发式、时段窗、等候费、末端拒收二次配、日清核对。触发:城配、多点、同城、日清、等通知、拒收。
Control OnMyAgent's built-in in-app browser for opening, navigating, inspecting visible or interactive page state, clicking, typing, screenshots, uploads, downloads, dialogs, and local web testing. Prefer purpose-built connectors or APIs for semantic operations when available.
Control OnMyAgent's built-in in-app browser for opening, navigating, inspecting visible or interactive page state, clicking, typing, screenshots, uploads, downloads, dialogs, and local web testing. Prefer purpose-built connectors or APIs for semantic operations when available.
Use when a user asks to create, read, edit, review, or inspect a DOCX or Word document through the bundled Documents artifact plugin.
Use when a user asks to create, read, merge, split, rotate, inspect, or edit PDF files through the bundled PDF artifact plugin.
Use when a user asks to inspect or change a workbook already open in Microsoft Excel through the bundled Spreadsheets artifact plugin.
Use when a user asks to create, read, clean, analyze, calculate, convert, or edit XLSX, XLS, CSV, or TSV files through the bundled Spreadsheets artifact plugin.
Create, read, edit, review, render, and verify local DOCX and Word-compatible document artifacts. Use for documents, docs, Word files, .docx, memos, reports, letters, templates, tracked changes, comments, or a local DOCX intended for later Google Docs import. Do not use when the durable target is a PDF or spreadsheet.
Inspect or edit a workbook that is already open in the Microsoft Excel desktop app, but only when the current OnMyAgent session exposes a real Excel, connected-document, or Computer Use tool surface. Do not use for standalone spreadsheet files; use spreadsheets instead.
Read, create, inspect, extract, edit, render, and verify PDF files when the durable input or output is a PDF and visual layout matters. Covers text and table extraction, merge, split, rotation, watermarking, forms, OCR, encryption, metadata, and image extraction.
Create, read, edit, analyze, recalculate, render, and verify standalone spreadsheet files, including .xlsx, .xls, .xlsm, .csv, and .tsv, plus local XLSX files intended for Google Sheets import. Use when the durable input or output is a workbook or spreadsheet file. Do not use this skill to control a workbook already open in the Microsoft Excel desktop app; use excel-live-control when that capability is actually available.
Control local macOS apps through OnMyAgent Computer Use. Use for tasks that require reading or operating native app UI, accessibility elements, windows, screenshots, keyboard input, scrolling, or background-safe clicks when no purpose-built connector, API, or CLI is a better fit.
OnMyAgent documentation audit workflow. Use when scanning project docs for stale commands, old branding, broken links, duplicated state, roadmap drift, generated report clutter, local .loop state rules, or when updating README, Architecture, BUILD, AGENTS, loop-rules, package README files, and project skills under .agents/skills.
Audit .agents/skills SKILL.md files for duplicates, overlapping triggers, stale paths, broken references, and missing frontmatter. Use when reviewing the project skill catalog, after adding/renaming skills, or on a periodic skills health check.
创业庄园知识模块·顺序判断。包含顺序框架、判断能力、关口识别。对应《创业可以学》第1-2章。被 manor-judge 调用,不直接面向用户触发。当用户问题涉及"顺序/判断/关口/为什么忙没结果/先后"时由 manor-judge 读取本 skill。
创业庄园庄主的接待与人设层。负责身份识别、新用户进庄、接待流程、灯塔五步前两步(识别+接住)。触发场景:(1)新用户第一次进庄 (2)用户打招呼/寒暄/自我介绍 (3)不确定对方身份时 (4)用户主动问"你是谁""这里是什么地方" (5)用户只聊天不涉及具体创业问题。与 manor-judge 互斥触发:本 skill 只负责接待和身份识别,不回答具体创业问题。
NeoData Financial Search — natural language financial data search. Query stocks (A-share/HK/US), funds, indices, sectors, macro economics, forex, commodities in natural language. Covers real-time quotes, financial statements, capital flows, analyst ratings, announcements. Use when user asks about stock prices, earnings reports, fund performance, market data, GDP/CPI, exchange rates, gold, futures, or any financial data query.