make-to-markdown：用 markitdown 把多格式文档统一成 RAG 可用的 Markdown
make-to-markdown（@indiv-ebandao/make-to-markdown，v1.0.11）是 SkillHub 知识管理类 Skill，用 markitdown 将 PDF、Office、表格、HTML、图片、音频、ZIP、YouTube、EPub 等转为 UTF-8 Markdown，并支持 .doc/.ppt 经 COM 或 LibreOffice 预处理。统一入口 scripts/convert.py 负责依赖安装、降级与后置清洗，批量场景有 Init-Step-Poll 与 V1–V6 校验。可通过 skillhub.md 指引或 skillhub install --dir 安装到 Cursor 等 skills 目录；需本机 uv 与 python。
Read MoreRAG 知识库搭建：企业级 RAG 从零到 Pipeline 的 Agent Skill
SkillHub 社区 Skill「RAG 知识库搭建」（@user_69009747/rag-knowledge-base-builder，v1.0.0）面向企业级 RAG 从零搭建：固定/语义/结构化分块与小切大优化，BGE、BGE-M3、OpenAI 嵌入选型（含 MTEB 参考），LangChain Pipeline（摄入、向量化、检索、Rerank、生成、Ragas 评估），向量库可选 ChromaDB、FAISS、Milvus，并提及一键部署脚本。文章说明 Skill 定位、核心能力，以及通过 skillhub CLI 或安装文档在 Cursor 等环境启用的方式。
Read More提示词大师-java版：把画面意图转成 AI 可执行的专业提示
提示词大师-java版是 SkillHub 社区 ai-agent 类 Skill（@user_5d00e27d/majia-system-prompt-java，v1.0.2），通过马甲系统将脑海中的画面整理为 AI 可理解的专业提示。首次会追问确认方向，后续借助缓存加快对话，并用 RAG 将成功案例入库以持续优化。标签 requires_api_key=true，需配置 API Key。安装可用 skillhub install 并指定 Cursor 等客户端的 skills 目录，或让 Agent 按官方安装文档操作；对话中输入「新手指南」查看完整说明。
Read Moredsh-obsidian-board：把本地 Obsidian 变成 DSH 的本地知识库对话面板
`dsh-obsidian-board` 是 DeepSeek Harness 的本地检索插件，将 Obsidian vault 变成 RAG 面板。它采用双半架构：宿主端负责 BM25 索引与检索，浏览器端负责 UI 交互。具备极快的增量索引速度、点击引用跳转原文、可随时停止生成、以及完全的本地安全隔离（历史存储在浏览器本地，优先检索自有笔记）等特点。安装后配置 vault 路径，即可实现针对笔记内容的问答与总结。
Read More本地知识库智能问答：用个人文档做问答、报告与向量索引
介绍 SkillHub 社区技能「本地知识库智能问答」（@user_4c392ef3/knowledge-qa，v1.0.0）。面向 PDF/Markdown/Word 等个人资料，支持基于知识库问答、生成报告与思维导图、上传索引与分区/向量库管理；触发词涵盖「基于知识库」「查一下知识库」「整理成报告」等。说明 skillhub CLI 安装到 Cursor 等 skills 目录的方法，并提示 requires_api_key 标签与安装路径注意事项。
Read MoreIMA知识库法律咨询专家：用 RAG 做法律问题初步诊断与行动策略
IMA知识库法律咨询专家是 SkillHub 社区 Skill（@user_b463487e/tencent-ima-copilot-legal-consultation，v1.1.2），以资深律师分阶段办案思维结合 RAG 检索，提供法律纠纷的初步诊断与行动策略。支持分层精读、多轮换词搜索、结构化引用与效力标注、快速/专家双模式，并内置绝对溯源、时效验证与幻觉自检。适用于欠钱不还、劳动、合同、家事等常见咨询及法律文件分析，不替代正式法律意见与出庭。requires_api_key=false，按 skillhub.cn/install/skillhub.md 安装即可。
Read More提示词大师-python版：把画面意图翻译成 AI 可用的专业提示词
提示词大师-python版是 SkillHub 社区 ai-agent 类 Skill（v1.0.5，@user_5d00e27d/majia-prompt-translator-python），通过马甲系统将脑海中的画面翻译成 AI 可用的专业表达。支持首次多轮澄清、缓存加速后续对话，以及成功案例入库的 RAG 进化。标签 requires_api_key=true，需按 skillhub.md 用 skillhub install 安装到 Cursor 等 Agent 的 skills 目录；对话中可输入「新手指南」查看完整说明。
Read Morememory-systems: A Design Guide for Agent Persistent Memory and Retrieval Architectures
memory-systems is the Memory System Design guide in the Agent Skills for Context Engineering collection, covering cross-session semantic memory, entity consistency, temporal validity, graph/vector and hybrid retrieval, and memory consolidation. The article explains its positioning, activation boundaries, memory hierarchy and framework selection table, and principles for consolidation and JIT loading. It also provides official examples using Mem0, temporal graphs, and Cognee, along with loading methods via Claude Code and npx skills, emphasizing starting with a shallow design, validating it with LoCoMo/LongMemEval, and then deepening the architecture.
Read MoreDeepSeek Harness Plugin: searchsieve — Search Result Coverage Assessment Tool
The DeepSeek Harness plugin `searchsieve` is designed to address the challenge of quantifying search result quality in RAG scenarios. It provides core metrics such as retention ratio and answer completeness, and performs semantic analysis and evaluation of search results. Using a pure rule-based algorithm, it runs with zero model overhead and requires no API Key, while being optimized for Chinese environments. To install it, run `dsh plugin add search-sieve` and restart after installation. The plugin validates information points through the `expectedPoints` parameter and returns the selected ratio, uncovered information, and follow-up search suggestions. On Windows, it can be used with `dsh.cmd` or `npm.cmd`.
Read MoreDeepSeek Harness Local Desktop Knowledge Base Plugin
`dsh-move-rag` is a Windows desktop plugin for DeepSeek Harness that requires no model weights or cloud services and supports local knowledge base management. Through a topmost WinForms panel, it allows users to drag and drop PDF, Word, image, and other files for rapid ingestion. The plugin uses hash embedding techniques to enable offline, fast local RAG retrieval and exposes the results as tools to Agents. This solution is suitable for Windows users who want to quickly build a local knowledge base, but images are indexed by file name only and the implementation relies on the desktop panel.
Read Moredsh-plugin-rag: DeepSeek Harness Semantic Memory Plugin
`dsh-plugin-rag` is a DeepSeek Harness plugin that uses RAG to convert chat sessions into searchable, local persistent memory. It features automatic indexing, incremental updates, and non-repudiation support, and allows retrieval of historical information through the `rag_search` tool. Installing it requires modifying `package.json` and `cordis.patch.yml`, after which it becomes available upon restart. Index data is stored by default in local JSON files, and restarts are idempotent.
Read Moredsh-plugin-rag: DeepSeek Harness Local Knowledge Base Retrieval Plugin
The DSH Harness framework provides the `dsh-plugin-rag` plugin, designed to address local codebase retrieval needs for agents while avoiding the latency, cost, and privacy risks associated with external APIs. The plugin supports indexing Markdown, code, and configuration files, using a purely local TF-IDF algorithm for keyword matching without additional dependencies. Once installed and enabled, it can search files locally or display index statistics through the `rag` command, making it suitable for agent scenarios that require decision-making based on local data.
Read Moresatan9394/dsh-rag: RAG (Retrieval-Augmented Generation) Skills under DeepSeek Harness
`dsh-rag` is a RAG skill plugin for DeepSeek Harness (DSH), designed to improve an agent's information accuracy and response quality through retrieval-augmented generation. The plugin provides core capabilities such as vector stores and embeddings, document pipelines, retrieval strategies, and hallucination control. Since the installation command is missing and no examples are provided, users need to check the GitHub repository themselves for functional details.
Read Moredsh-kb-rag: A Local-First RAG Knowledge Base Toolkit
The DSH plugin `yumu247/dsh-kb-rag` provides a local-first RAG knowledge base toolset for DSH, designed for zero-cost, local deployment. The plugin wraps four core commands: `kb_query` for semantic retrieval, `kb_ingest` for incremental ingestion, `kb_crawl` for web scraping using Scrapling, and `kb_related` for graph-based related recommendations. Before installation, ensure the host is configured with a Python environment, an Ollama model (such as `bge-m3`), and the `kb-rag` pipeline. During installation, specify the `kbRagDir` path, which the plugin relies on to run. When in use, `kb_query` returns results in JSON format, and `kb_inge
Read Moredsh-skill-indexer: A Two-Stage Intent Retrieval Plugin for Local SKILL.md
dsh-skill-indexer is a Host Bundle plugin for DSH, designed to address the challenge agents face in precisely locating local SKILL.md files. It establishes a two-level indexing mechanism that leverages keyword coverage and IDF for intent routing, supporting a three-state fallback (hit/low/miss). The tool runs entirely locally with zero dependencies, offers both command-line and programmatic invocation, and supports hash-based incremental updates.
Read Moredsh-understand-anything: A DeepSeek Harness Plugin for Generating Codebase Knowledge Graphs
`dsh-understand-anything` is the codebase analysis plugin for DeepSeek Harness. By generating structured knowledge graphs, it addresses the issues of traditional RAG consuming excessive amounts of tokens and easily losing architectural details. The plugin supports architecture understanding, change impact analysis, onboarding guides, and multi-domain extraction without requiring raw files to be stuffed into the context. Installation requires Node.js ≥ 22 and pnpm ≥ 10. Please note that, due to plugin compatibility limitations, Web applications, Claude Code–specific features, and the analysis engine have not been translated.
Read Moredsh-doc-compiled-skills: Pre-compile documents into agent-executable skills
This article introduces the DSH community skill "satan9394/dsh-doc-compiled-skills". Its core approach is to pre-compile long documents into executable skills during the build phase, rather than performing RAG retrieval at runtime, thereby making agent behavior more deterministic and reducing context overhead.
Read Moreoh-my-knowledge: Verifiable Knowledge Change Assessment for Prompts, RAG, and Agents
OMK (oh-my-knowledge) is a MIT-licensed CLI and DSH plugin maintained by lizhiyao, designed for knowledge maintainers working with prompts, RAG, skills, agents, and workflows. Through a closed loop of doctor, eval, and observe, it compares versions of knowledge artifacts under the same model and samples, outputting verifiable release conclusions, failed samples, and costs. It also supports viewing DSH task traces in Studio. DSH users can install it in the web profile using `dsh plugin --profile web add oh-my-knowledge` and execute `/omk eval` and `/omk observe` within a session.
Read More