@tianheihei002

mj-gallery automates midjourney image generation through polling, local download, gallery archival, and web deployment. enforces english prompts, aspect ratio selection, and returns both local paths and public gallery links.

yq-image-creator routes users across 17 curated image generation styles spanning characters, scenes, products, and style transforms. uses form wizards for selection and delivers outputs via xml asset tags with minimal interaction overhead.

yq-feishu-webhook-skill sends feishu webhook messages in plain text, rich text (post), and interactive card formats. supports image upload via tenant access token. comprehensive reference material provided but procedure clarity and error handling need strengthening.

html-presentation-generator creates multi-slide decks with typed layouts (cover, toc, divider, content, summary), mandatory image generation, and step-by-step workflows. includes extensive color palettes and design system reference but lacks error handling and recovery strategies.

yq-find-skills provides skill discovery and lifecycle management for agent teams, including missing skill identification, semantic search, and compatibility checking. execution remains vague on actual matching algorithms and integration points with the underlying skill graph.

yq-industry-research-report executes a 4-phase workflow (research, writing, fact-checking, formatting) to produce data-driven industry reports in multiple formats. emphasis on multilingual sourcing and strict procedural compliance, but workflow clarity deteriorates significantly in later phases and lacks recovery mechanisms for common failure modes.

yq-skill-sync detects and synchronizes skill versions across a clawhub workspace via cli integration and version api polling, with daily cron scheduling and local skill manifest scanning.

blueprint-generator converts natural language and screenshots into unreal engine blueprint and material node graphs with ai-powered visual editing, but lacks actionable procedures and outcome validation.

yq-ppt-master covers ppt creation, modification, and design consultation with emphasis on minimalist aesthetics. procedural workflows exist but lack concrete decision criteria, error handling, and testability signals.

yq-story-video-generator accepts flexible inputs (1-N images, text, or mixed) to produce multi-segment story videos with automatic script generation, character consistency via reference images, and background music synthesis. workflow is detailed but untested in production.

comfyui-workflow-generator converts natural language descriptions into comfyui json workflows via a three-stage pipeline (generate, validate, build), supporting multiple node types but lacking detailed procedural steps and explicit error handling.

yq-knowledge-digest converts learning materials into multimodal outputs (notes, quizzes, slides, mindmaps, audio). the skill defines desired formats and outputs clearly but lacks actionable implementation steps, decision logic, and failure handling.

大师级AI提示词优化专家。当用户需要优化提示词、改进prompt、进行提示词工程、设计高效提示词时使用此技能。将模糊的请求转化为精准构建的提示词，最大化释放AI在所有平台的潜力。Keywords: prompt optimization, prompt engineering, 提示词优化, 提示词工程, imp...

专业 AI 视频创作技能，帮助用户使用 Seedance 2.0 模型创作高质量视频内容。当用户需要生成 AI 视频、编写视频提示词、使用即梦平台、进行多模态视频创作、制作分镜、设计视觉风格、优化提示词时调用。支持图像、视频、音频、文本四种模态输入的提示词优化和创作方案设计，整合专业电影制作知识体系。

专业的文件夹整理助手，帮助用户安全地整理和清理文件夹。在清理前自动创建压缩备份，使用移动命令代替删除命令确保数据安全。支持Mac/Linux和Windows系统，需在桌面端使用。触发关键词：文件整理、文件夹清理、清理文件、整理目录、folder cleanup、organize files

mxai 创作助手，聚合多种 AI 图片和视频生成模型，以及 MJ（MidJourney）绘画。关键词：AI绘画、图片生成、视频生成、AI创作、mxai、MJ、MidJourney

Generate Unreal Engine material node graphs from natural language descriptions, supporting key nodes like TextureSample, Multiply, Lerp, and MaterialOutput.

雨晴定时任务系统管理器。当需要检查cron任务状态、修改任务配置、排查重复职责、优化任务安排时触发此技能。也是主人确认的"主动发现+解决问题"能力核心。

哔哩哔哩CLI工具，支持抓取视频详情/字幕/AI摘要/评论/相关视频，提取音频并分割为ASR-ready WAV片段，获取用户资料/视频列表/关注列表，搜索用户或视频，热门视频/全站排行榜，动态时间线，收藏夹/稍后观看/观看历史，点赞/投币/一键三连等互动操作。当用户要求获取B站视频信息、下载音频、查询UP主信息...

职业规划未来镜像系统：收集用户信息，生成3条差异化职业路径的Awwwards级HTML报告，构建'3年后的自己'进行镜像对话，生成《来自未来的信》沉浸式HTML页面。触发关键词：职业规划、职业转型、职业困惑、未来镜像、职业蓝图

Q版表情包生成器：将用户照片（人物、宠物、物品、Logo/Icon）转换为4个经典动作的Q版GIF动态表情包，带中文配文，支持在线展示和下载。关键词：表情包、sticker、GIF、Q版、动图

热点追踪专家：基于用户输入的主题，搜索最新信源、挖掘近期爆点话题，并生成高质量长文内容。适合追踪热点事件、行业动态、人物资讯等持续性话题。关键词：热点、追踪、爆点、话题、长文、新媒体、信源