画音灵匠：把国风想法翻译成 Suno 可用的双框提示词
画音灵匠是 SkillHub 社区 Skill @user_1793b15c/ai-music-prompt（v1.0.2，design-media），由七仔的AI工具箱发布，无需 API Key。它将国风写词、pypinyin 格律辅助、spec 结构化设定与 Suno 的 Style of Music、带段落标签的 Lyrics 组装、finetune 人话微调及 diagnose 改歌诊断串成「写词→组装→试听→微调」流程，附带短语库与 Suno 风格词表。本文说明定位、核心脚本、skillhub 安装路径与典型命令，并注明其不直接生成音频、效果依赖 Suno 平台。
Read More东方仙侠视觉导演：在统一原典下生成与诊断仙侠场景提示词
东方仙侠视觉导演是 SkillHub 社区 Skill（@user_11893326/xianxia-visual-director，v1.0.0，design-media），在继承式主原典下生成、派生、优化并诊断明亮电影感东方仙侠环境提示词，强调可度量神性尺度、五层空间、40–60% 呼吸留白、纪念碑式东方建筑、无栏杆边缘及仙人与服饰一致。覆盖云上仙城、仙界大境原典等路线，支持多种画幅且单条默认 16:9。requires_api_key=false。安装请按 skillhub.md 安装该坐标。
Read MoreAI提示词设计服务：把提示词从「能用」做到稳定可控
介绍 SkillHub 社区 Skill「AI提示词设计服务 - 全行业专业提示词优化工程」（@user_9e8820a0/ai-prompt-designer，v1.0.0，ai-agent）。涵盖模板、定制、深度优化、风格调整与拆分改写，无需 API Key。说明通过 skillhub CLI 安装到 Cursor、Claude Code、Codex 等 skills 目录的步骤，以及基于官方简介的典型使用方式与适用注意。
Read More图生视频提示词：为 Image-to-Video 模型生成中英双语运镜提示
SkillHub 社区技能「图生视频提示词」（@user_5d79bc35/videocreate-19921119，v1.0.0）基于 aiandcivilization/Img2Vid-Prompts，为 ImagenVideo、Runway Gen-2/Gen-3、Pika、可灵、即梦等图生视频模型生成结构化提示词，强调电影运镜与视觉特效术语。每次生成强制输出中英文双语：英文用于 Runway/Pika/Sora，中文用于可灵/即梦/通义。安装需参考 skillhub.cn/install/skillhub.md，用 skillhub install 并 --dir 指向 Cursor 等 Agent 的 skills 目录；无需 API Key。
Read More图像提示词工程师 image-prompt-engineer：把视觉概念写成可复用的文生图提示词
image-prompt-engineer 是 SkillHub 上由 bozoyan（@indiv-bozoyan/image-prompt-engineer，v1.0.0）发布的文生图提示词 Skill，分类 design-media。它将画面拆为主体、环境、构图、光线、风格、画质六层，配套负向提示、比例与留白规则，并输出中英双版及 Midjourney、SDXL/Flux、即梦等写法差异。通过 skillhub install 安装到 Cursor、Claude Code 或 Codex 的 skills 目录后，在对话中用「图片提示词」等触发语即可按 SKILL.md 模板生成；另可选按次付费 HTTP 接口。本文基于 SkillHub 页面与 SKILL.md 整理安装步骤与典型用法。
Read MoreStyle Design Generator：用六槽位基因库生成可复用的风格提示词方案
Style Design Generator 是 SkillHub 上的 design-media 类 Agent Skill，作者王教成（波动几何），基于《字体设计提示词方案》与《信息图设计提示词方案》各 100 例，用六槽位模板提供换主体、要素混搭、模板衍生、新风格延展与领域应用五类能力。安装命令为 npx secure-skills-cli add style-design-generator，可用 skillhub verify 校验签名。适合需要可编号、可复用视觉风格提示词与跨媒介映射的创作者；使用时宜分步触发并核对槽位协调性。
Read More提示词专家：把含糊需求落成可测、可迭代的提示词
SkillHub 社区 Skill「提示词专家」（@user_8d26dabd/qclaw-prompt-engineer，v2.1.4）面向提示词输出不稳定、需求边界不清的场景，将含糊目标整理为更可执行的提示词，并提供约 5–10 条测试用例与迭代方法以对齐预期。无需 API Key（SkillHub 标注）。本文基于 SkillHub 页面与安装文档说明 skillhub CLI 安装步骤，以及 Cursor、Claude Code、Codex 等常用 skills 目录示例，并归纳适用场景与使用注意。
Read Morecoding-prompt：在写码会话里把「给 AI 的指令」写清楚
coding-prompt 是 SkillHub 社区 Skill（@user_5b28ea14/coding-prompt，v2.0.0），定位为 AI coding prompt optimizer and coach，面向开功能、纠偏、审代码、要测试等写码场景，帮助补全约束与验收标准并优化编程向 prompt。无需 API Key。可通过 skillhub CLI 安装到 Cursor、Claude Code、Codex 等 Agent 的 skills 目录。非编程闲聊不应触发。
Read More提示词大师-CLI版：把画面翻译成 AI 能执行的专业提示词
提示词大师-CLI版是 SkillHub 社区 ai-agent 类 Skill（@user_5d00e27d/majia-system-prompt-cli，v1.0.1），通过马甲系统将口语化画面描述整理为 AI 可执行的专业提示词，并强调缓存与 RAG 让越用越快、越贴习惯。文章说明其定位、核心能力、SkillHub CLI 安装路径（含 Cursor/Codex 等 skills 目录）及输入「新手指南」的典型用法，并注明 requires_api_key 与社区来源注意事项。
Read MoreSeedance 2.0 电影级 AI 视频提示词专业版：用 Skill 把提示词写成可复用流程
本文介绍 SkillHub 社区 Skill「Seedance 2.0 电影级AI视频提示词-专业版（全域完整版V1.0）」（v1.1.0，@user_2df7b78c/seedance2-prompt-expert）。它基于 SKILL.md，整合五大心智模型、四层建筑法、@ 多模态绑定优先级、八大行业模板、20+ 场景诊断表与能力边界，并定义 Agent 触发词与标准入参/出参，无需 API Key。文中给出 skillhub CLI 安装步骤及 Cursor、Claude Code、Codex 的 skills 目录示例，说明从安装到生成与瑕疵诊断的典型用法与注意事项。
Read More即梦-文生图提示词技能：为即梦整理、优化文生图 Prompt
SkillHub 社区 Skill「即梦-文生图提示词技能」面向 Dreamina/即梦文生图，将 35 类场景、115+ 注释范例、60+ 主题词库及中国传统色、故宫 384 色与 Pantone/RAL 配色写进 SKILL.md，供 Agent 在用户写、改、翻译画面描述时产出可上线 Prompt。坐标 @user_c1d16043/jimeng-prompt-text2image，v1.0.1，无需 API Key。安装需 skillhub CLI 并用 --dir 指向 Cursor 等客户端 skills 目录。
Read More即梦-图生图提示词技能：用 Keep/Change 把图生图提示词写准
SkillHub 社区技能「即梦-图生图提示词技能」面向即梦 4.0+ 图生图编辑，用 Keep/Change 明确保留与变更项，覆盖风格迁移、换背景、细节增强、色彩调整、元素增删、季节时间、修复、换装、创意变形与多图合成等十类场景，含 30 个注释示例与 I2I 模板。无需 API Key。可通过 skillhub CLI 安装到 Cursor、Claude Code、Codex 等 skills 目录，或由 Agent 按官方安装文档加载。
Read MoreAI绘图大师：用 SOCAWT 框架做文生图提示词工程
AI绘图大师是 SkillHub 社区技能 @user_19bbb2a6/imageprompt-socawt（v1.0.1，design-media），基于 SOCAWT 六维视觉分析框架，用自然语言长句而非标签堆砌，为即梦、GPT-Image-2、Nano Banana 等模型编写提示词。支持灵感创作（纯文字）与逆向提取（参考图、先识图类型再自适应分析）。requires_api_key=false。安装需按 skillhub.cn/install/skillhub.md 使用 skillhub install 并指定 Agent 的 skills 目录。
Read More玄奘.skill：在用户施压或反复失败时切换 Agent 工作姿态
玄奘.skill 是 SkillHub ai-agent 分类下的提示词 Skill，在用户明确请求 pua 模式、反复失败两次及以上、抱怨质量或发送 /pua 等触发词时切换 Agent 策略，正常首次编程与查询则不触发。本文基于 SkillHub 公开简介与安装文档，说明触发边界、skillhub CLI 与 Cursor 等 skills 目录安装方式，以及典型对话用法与误触发注意。官方入口：https://api.skillhub.cn/user_741dc82b/xuanzang-skill。
Read MorePrompt Generation Master: A Unified Multimodal Prompt Skill for Images and Videos
Prompt Generation Master (@user_5d00e27d/image-prompt-skill, v1.0.15) is a SkillHub community skill designed as a unified prompt generator supporting both image and video modality: text-to-image and image-to-image, text-to-video and image-to-video, image/video decompilation, and context fine-tuning. It claims automatic adaptation to art styles and platform syntax, with built-in RAG-based style presets and safety validation; requires_api_key=false. Based on the official SkillHub introduction and installation documentation, this article outlines its capability boundaries, the skillhub CLI installation command, and the skills directories for different Agents, providing a reproducible usage workflow and precautions.
Read MoreLuBan.skill: Scoring and Automatic Optimization for SKILL.md
Luban.skill (luban-skill-pro, v1.0.6) is an Agent skill optimizer on SkillHub that provides ten-dimension rubric scoring, quality checks, and iterative automatic optimization for SKILL.md. It supports Quick mode (lightweight multi-round refinement with .bak rollback) and Full mode (git branches, multiple reviewers, and strict review), while incorporating human confirmation at key checkpoints. This article explains its positioning, core capabilities, steps for installing it via skillhub.cn into the skills directories of clients such as Cursor, as well as typical natural-language-triggered usage and precautions.
Read MoreDeepSeek Harness Plugin: dsh-prompt-workbench
The DeepSeek Harness plugin `dsh-prompt-workbench` provides a prompt rewriting workbench designed to address the issue of agents deviating from intent due to ambiguous input. The tool runs locally in the browser, supports prompt analysis across six dimensions, including goals and context, and offers four rewriting modes, such as refining requirements and adding constraints. The plugin features streaming output, token-level diff comparison, structured suggestions, and dual-interface synchronization, enabling developers to transform ambiguous input into verifiable and executable structured prompts.
Read MoreDeepSeek Harness Plugin: dsh-prompt-library Common Prompt Library
`dsh-prompt-library` is a locally running plugin designed to lower the barrier to using DeepSeek Harness. It provides more than 30 Chinese prompts covering multiple scenarios, supports local search and one-click insertion, and processes all data entirely in the browser. Users can click the prompt library button, search for and insert a template, replace the placeholders, and then send the prompt. The plugin requires no internet connection and is suitable for users who are not familiar with prompt engineering or who want to quickly access commonly used templates.
Read Moredsh-prompt-studio: DSH Prompt Studio
`dsh-prompt-studio` is a DeepSeek Harness (DSH) plugin built to address inefficient prompt iteration and poor traceability. It provides a unified prompt workspace in the left sidebar, with core features including prompt optimization and rewriting, multi-round iteration, side-by-side version comparison, and template and history management. It supports using optimized prompts directly in conversations and provides a `prompt_optimize` tool for agent invocation. After installation, DSH must be reloaded; data is stored in a local SQLite database.
Read MoreDeepSeek Harness (DSH) Prompt Optimizer Plugin
The DeepSeek Harness plugin `co1ombiagly/prompt_optimize_dsh` aims to address the high barrier to entry in prompt engineering. It can convert casual requirements into rigorously structured prompts with one click and automatically insert them into the input field. Featuring context awareness, iterative rewriting, and zero API key management, it ensures data security. Users can quickly generate high-quality prompts with the `/optimize` command, without needing to manage an API Key.
Read MoreDeepSeek Harness Plugin: PromptKit
PromptKit is a plugin for DeepSeek Harness designed to turn rough drafts into structured, actionable prompts. It includes the PromptStudio advanced workspace and QuickEnhancer conversational quick enhancer, supporting one-click enhancement, five-dimensional diagnosis, a library of 22 thinking methods, local management, and version control. The plugin features zero telemetry and a zero-token option, supports offline work, and helps developers efficiently build and optimize prompts through capabilities such as streaming output and automatic polishing.
Read Moredsh-prompt-engineering: DeepSeek Harness Prompt Engineering Patterns
The `dsh-prompt-engineering` plugin, developed by `satan9394`, is an extension for DeepSeek Harness designed to streamline prompt creation. It provides standardized prompt engineering patterns, supports CoT, few-shot examples, and template reuse, and helps improve debugging efficiency. Since no official installation command is provided, users should refer to the community directory or GitHub for guidance. During installation, proper permissions and security compliance must be considered.
Read Moredsh-gpt-file-discipline: A Routing-Level Prompt Assembly Plugin for DeepSeek Harness
The DeepSeek Harness plugin addresses the cumbersome problem of dynamically adjusting system prompts when an agent switches across multi-model routes. It automatically switches prompts based on routing conditions: it adds a file-reading discipline for GPT models and appends a designated thinking language for DeepSeek routes. It supports exact matching, on-demand registration with no additional overhead, and allows customizing text and section names via configuration. Once installed, it automatically matches GPT models under the codex provider or models prefixed with deepseek, improving prompt management efficiency in mixed-model scenarios.
Read Moresatan9394/dsh-ai-image-design: AI Image Generation Workflow and Batch Prompt Derivation
The DeepSeek Harness (DSH) plugin satan9394/dsh-ai-image-design focuses on image generation, providing AI image generation workflows and batch prompt derivation capabilities. As a plugin that runs with DSH process privileges, it is suitable for handling specific image task workflows. It currently lacks official installation commands, usage examples, and clear license information, so the source code should be carefully reviewed before use.
Read Moredsh-media-gen: Transforming Media Briefs into Production-Ready Prompts and Atlas Cloud Workflows
`dsh-media-gen` is a DeepSeek Harness (DSH) plugin maintained by AtlasCloudAI, designed to simplify multimedia workflow development under the "everything is a plugin" philosophy. It automatically converts media briefs into model selections, prompts, and workflow solutions, supporting workflows for images, video, audio, and 3D content. The plugin addresses pain points such as model selection, video planning, and prompt generation, and provides MCP bridging with Atlas Cloud to enable automated execution. After installing the plugin and configuring an API key, users can handle media generation requests directly within DSH. The project is released under the MIT license.
Read Moredsh-ctm: Turning Model Context into First-Class, Visualizable, Editable, and Scored Objects
`dsh-ctm` is a transparent context management plugin for DeepSeek Harness, designed to address the difficulty of controlling mixed and hard-to-manage context information. Built with TypeScript, it provides context segmentation visualization, MECE token accounting, real-time editing, and effectiveness scoring. It supports modifying the initial prompt, intercepting and altering the context stream, and includes a built-in scoring engine to assess content quality. This tool enables a closed loop of “view, edit, and evaluate” for context management, making it well suited for developers who need fine-grained debugging and prompt optimization.
Read More