Agent Skills

---

name: cpa-diagnostics

---

name: creative-fatigue-detection

---

name: day-hour-performance-breakdown

---

name: device-performance-split

---

name: e2e-seo-assistant

---

name: email-sequence-writer

---

name: frequency-cap-recommendations

---

name: geo-performance-analysis

---

name: adhd

description: Parallel divergent ideation for coding agents. Spawns N isolated branches under different cognitive frames (regulator, biology, speedrunner, 10-year-old, $0 budget), scores, clusters, prunes traps, and deepens top survivors. Use on /adhd, "ADHD mode", brainstorm/ideate intents, or open-ended design, architecture, naming, API/SDK surface, and fuzzy-debugging decisions. Skip for syntax, lookups, bugs with known root cause, or closed phrasing ("quick", "standard", "canonical", "textbook"). Full pre-flight gate is in the skill body.

---

name: observal

description: Use when starting any task the organization may already have an approved skill, prompt, MCP server, or Agent for: reviewing code, a commit, a diff, or a pull request; writing tests or documentation; querying a database, API, or service; automating a browser or web page; running untrusted code; connecting to a SaaS tool; drafting, researching, or any other substantive work. The user will not mention Observal: the task type is the trigger. Check what is already installed, then run observal discover search as the first action, before reading the repository or working from scratch. Also use when the user wants to log in, configure Observal, inspect local harness setup, share or open a repository Agent share, manage a teamspace or invitation, process inbox items, check installed registry items, or call an endpoint without a dedicated command.

---

name: observal-admin

description: Administers Observal users, settings, diagnostics, review queues, security events, audit logs, SAML, SCIM, the local Observal server, its upgrades and rollback, and its own PostgreSQL and ClickHouse schema migrations. Use when the user needs privileged governance, submission decisions, identity configuration, security investigation, or Observal server operations. Not for querying a database or any work outside Observal administration.

---

name: observal-advanced

description: Recovers Observal session ingestion, manages CLI upgrades, downgrades and rollback, and performs explicit local Agent fallback when the server is unavailable. Use when the user asks to reconcile missed sessions, repair CLI version state, or continue locally after a confirmed connection or configuration failure.

---

name: observal-agents

description: Creates, forks, authors, validates, publishes, updates, versions, pulls, archives, restores, transfers, and manages co-authors for Observal Agents. Use when the user wants to build or install an Agent, change an Agent definition, publish a draft, release a version, or manage Agent ownership.

---

name: observal-ops

description: Inspects Observal traces, sessions, rankings, feedback, telemetry health, logs, and Agent insight reports. Use when the user wants operational evidence, current activity, telemetry diagnosis, ratings, report generation, regression analysis, or recommendations grounded in Agent usage.

---

name: observal-registry

---

name: x-article-draft-uploader

description: 将 Obsidian 或本地 Markdown 文章上传到 X/Twitter Articles 草稿；支持 Markdown pipe 表格转成 X 原生表格块；如果文章最上方有图片，则按当前 X 界面提示的 5:2 封面图上传；如果没有封面图，则跳过封面区域，直接上传标题、正文、表格和正文图片，完成后提醒用户补一张 5:2 封面。X Article 当前实测正文最多 25 个媒体项，封面单独上传且不占正文名额；表格尺寸选择器当前最大 10 x 10。适用于用户要求上传、发布、保存 Markdown 到 X Article，尤其是需要复用 Chrome 登录态、使用独立 Playwright 浏览器、不接管用户当前浏览器，或旧脚本出现缺图、错位、MPH_MARKER、表格变纯文本等问题时。

---

name: yichen-agent-memory

description: 安装、升级、检索或维护 Agent Memory Vault 共享记忆系统；用于明确的记忆系统任务。普通业务的历史回忆按 Vault 入口读取，不启动安装或维护流程。

---

name: yichen-asr

description: 逸尘自用的统一音视频转写入口，在 StepFun Step ASR 与火山引擎豆包 ASR 之间按输出需求、安全边界和可用状态路由。用于本地音频或视频的纯文本转写、时间戳、SRT 字幕、口播粗剪，以及转写前体检；用户明确指定服务商时不得静默切换。Use when a local audio or video file needs transcription and the correct ASR backend is unclear, or when the user explicitly invokes $yichen-asr.

---

name: yichen-bookmarks-export

description: 在用户当轮明确授权后，只读导出其小红书收藏、抖音收藏或 X/Twitter 书签为本地链接文件；本 Skill 直接完成收藏页滚动、链接提取、去重、数量核验、Field Theory GraphQL-only 导出和抽样验证。用于“导出/备份/同步我的收藏或书签链接”等私人数据导出任务；不得用于公开搜索、收藏内容分析、媒体下载、正文归档、推荐发现或未获当轮授权的私人读取。

---

name: yichen-chatgpt-web-research

description: Use the user's already signed-in official ChatGPT website account, especially GPT-5.5 Pro / ChatGPT Pro, to perform product research, market research, competitor research, second-opinion analysis, or any request that says ChatGPT 官网, ChatGPT 网页版, GPT-5.5 Pro, 5.5 Pro, 用官网调研, 产品调研, 市场调研, or avoid API cost. Prefer the Chrome plugin/extension route; if it is unavailable or cannot attach, fall back to Computer Use on the real visible Chrome window/profile. This skill must control the real ChatGPT web page in the intended Chrome account/profile; it must not open a different free-account window, use OpenAI API, use web search alone, use generic Playwright, or use the deprecated chatgpt-web MCP.

---

name: yichen-codex-chatgpt

description: Use the signed-in ChatGPT website in Chat UI Pro mode, never Work mode, to research, architect, and review one local project through a configured Secure Tunnel and read-only MCP, while Codex alone edits and tests locally. Trigger on explicit $yichen-codex-chatgpt, 让 ChatGPT 设计后由 Codex 写代码, ChatGPT 架构并审查 Codex, or requests for a Codex and ChatGPT collaboration loop. Pure research stays on the official ChatGPT website; code and hybrid work follow ChatGPT PLAN to Codex execution to ChatGPT REVIEW. Website control defaults to in-app browser, then Chrome, then bounded Computer Use. Use a compatible local-only MCP only when the user explicitly requests local-only mode and that optional dependency is configured, never as an automatic fallback.

---

name: yichen-content-archive

description: 按用户当轮要求读取、下载或归档已知网页/社交媒体链接、URL 清单、已确认候选或明确的有限容器。支持显式有界站点归档；不用于关键词发现、推荐扩展或私人收藏导出。

---

name: yichen-grok-consult

description: Consult or search with xAI Grok from a GPT-led Codex conversation without switching the main model. Use when the user asks Grok to answer, review, challenge, compare, provide a second opinion, search the web, search X/Twitter, find recent or fixed-date posts, decode timestamps embedded in X status IDs, research current topics, or invoke Grok while staying in the current GPT conversation.

---

name: yichen-jianying-edit

description: 在装有匹配版本剪映的 Apple Silicon Mac 上，无界面生成可编辑草稿、修改已有多轨草稿的独立副本，或按明确要求导出原生 MP4。需要另行检出 Jianying Headless 核心项目；纯转写不使用本 Skill。