CommunityCoding & Developmentgithub.com

conardli/kb-retriever

A retrieval and Q&A assistant for local knowledge base directories. Core process: (1) hierarchical index navigation, (2) when encountering PDF/Excel files, must first read references to learn the processing method, (3) process files then retrieve. Uses grep, Read, pdfplumber, and pandas in combination based on file type for progressive retrieval to avoid loading entire files.

What is kb-retriever?

The conardli/kb-retriever skill is designed to help AI agents (like Claude Code, Cursor, Codex, etc.) efficiently retrieve information from local knowledge base directories. Its core workflow involves three steps: first, it navigates a hierarchical index to understand the directory structure; second, when it encounters PDF or Excel files, it must read the `references` documentation to learn the correct processing method before handling the file; third, it processes the files and performs the actual retrieval. To optimize performance and avoid loading entire files into memory, the skill uses a progressive retrieval strategy that combines tools like `grep`, `Read`, `pdfplumber`, and `pandas`, selecting the appropriate tool based on the file type. This skill is particularly useful for developers or AI agent users who need to answer questions, search for information, or look up data within a local knowledge base directory, ensuring efficient and accurate retrieval without overwhelming system resources.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/conardli/garden-skills/tree/main/skills/kb-retriever

Installed? Explore more Coding & Development skills: steipete/bluebubbles, steipete/eightctl, steipete/blucli · View all 6 →

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

What does kb-retriever do?

A retrieval and Q&A assistant for local knowledge base directories. Core process: (1) hierarchical index navigation, (2) when encountering PDF/Excel files, must first read references to learn the processing method, (3) process files then retrieve. Uses grep, Read, pdfplumber, and pandas in combination based on file type for progressive retrieval to avoid loading entire files. Use this skill when the user's question involves "answering questions from a knowledge base directory / retrieving information / looking up materials."

Individual skills in this repo

This repo contains 3 individual skills — each has its own dedicated page.

conardli/gpt-image-2

面向 GPT Image 2 的图像生成 / 编辑技能。可在 3 种环境下使用:(A) Garden 本地模式,通过 OpenAI 兼容接口直接出图并落盘;(B) Host-Native 模式,把本 Skill 当作提示词工程指引,把渲染好的 prompt 交给宿主 Agent 自带的图像工具出图;(C) Advisor 模式,宿主无任何图像工具时退化为高质量 prompt 顾问。涵盖 18 大类、80+ 个结构化模板,覆盖海报 / UI / 产品 / 信息图 / 学术图 / 技术架构图 / 漫画 / 头像 / 流程板 / 电影分镜 / IP 周边 / 编辑工作流等场景。

conardli/web-design-engineer

Build polished visual web artifacts with HTML/CSS/JavaScript/React: pages, dashboards, prototypes, slide decks, animations, UI mockups, and data visualizations. Use when the user wants a browser-rendered, interactive, or presentational front-end deliverable. Not for back-end, CLI, or non-visual coding tasks.

conardli/web-video-presentation

把一篇文章或口播稿,做成"看起来像视频"的点击驱动 16:9 网页演示,可选合成口播音频。流程:原始文章 → **一次产出**口播稿 + outline 开发计划 → 用户**一次对齐** 5 件事(稿子 / outline / 主题 / 素材 / 开发模式)→ 网页开发(逐章 / 顺序 / 并行)→ 可选音频合成(provider-agnostic:内置 MiniMax mmx-cli + OpenAI TTS,可换 ElevenLabs / edge-tts / Azure / 自带 TTS)。**outline 只规划节奏与信息密度,不规划动画** —— 动画由章节开发时按 PRINCIPLES + ANTI-AI 法则即时设计。每次点击推进口播稿的一个节拍,每一步独占整屏,进度条平时隐藏只在悬浮时出现。适用场景:用网页做视频(动态 PPT 但不像 PPT)、把口播稿 / 文章变成可交互的解说、为 B 站 / YouTube / 视频号录屏教程、做有电影感的产品 / talk demo。本 Skill 沉淀的是设计方法论 + 协作流程 —— 不绑定任何特定样式 / 字体 / 颜色 —— 因此能复用到任意主题与美学。

Related Skills