Recently Added
Fresh skills just added to the collection. Be the first to try them out!
harness-kit
by DanielWillson
Set up a project the Harness Kit way: machine + repo safety floor, tailored settings, a CLAUDE.md contract, a health-check audit, a knowledge wiki, a human README, and the building principles. Two flows: KICKOFF for a brand-new project ("run the kickoff", "kick off this project", project scaffolding) and ADOPTION for retrofitting the same harness onto an existing codebase ("adopt the kit", "retrofit the harness", "bring this project up to the kit's standards"). NOT for ordinary day-to-day work in a project that already has the harness.
浣跨敤IMYAI
by vegetpig
"浣跨敤IMYAI. Let Codex App call IMYAI official chat and image models by human-readable model name while Codex App remains the orchestrator and tool executor. Use when the user says 浣跨敤IMYAI, asks Codex App to use an IMYAI model for later replies, names an IMYAI model such as Claude Opus/Sonnet, Qwen, Gemini, Ava, GPT Image, Nano Banana, or asks for IMYAI image generation, login, JWT verification, model discovery, session mode, or IMYAI-first repair."
clawseccheck
by gl0di
Free, local security self-audit for your own OpenClaw agent. Reads your OpenClaw config, bootstrap files, log files, agent session logs, and installed skills — all read-only, all on your machine. Scores your setup (A–F) and reports the most urgent holes — reports only, it never changes anything. No API key, no data leaves your machine.
finance-trading-skill
by qingsuluan-dotcom
中文金融投研、量化分析与交易策略回测 Skill。Use when Codex needs to turn natural-language finance requests into reproducible research workflows, including stock or ETF trend analysis, financial statement review, industry research, screening, strategy design, historical backtesting, parameter scans, risk review, report generation, and data export. Supports A-share-first workflows with Tushare, AKShare, iWenCai-style search, Backtrader-style event backtests, VectorBT-style vectorized tests, and strict no-live-trading safety boundaries.
n0-app
by Lunar-Rails
Analyze a repository and generate a n0-app.json manifest and Dockerfile to make it deployable on N0's hosted apps platform. Trigger on: "make this deployable", "create app manifest", "n0 app", "deploy this on n0", "containerize this", "generate Dockerfile".
human
by Arif-2747
Use this skill whenever the user wants to humanize text, remove AI writing patterns, make prose sound natural, fix robotic writing, or write new content in their own voice. Triggers on: "this sounds like AI," "make this more human," "de-AI this," "rewrite this naturally," "write in my style," "remove the AI tells," or any request to edit prose for naturalness. Applies 38 documented patterns (29 from blader/humanizer + 9 stop-slop structural rules), voice injection, Style Mirroring (rewrite or write-from-scratch in the user's style from an uploaded sample), and Karpathy editorial discipline. Preserves full complexity, depth, and length. Never simplifies or summarizes.
gizza-tools
by gizza-ai
"Run gizza's local compute tools (calculator, text, image/video, fetch, …) from the terminal via the gizza CLI. The tool set is live — discover it with gizza list; never hardcode it."
show-and-tell
by wan-huiyan
Turn dense technical findings, an analysis, a code/research session outcome, or a jargon-heavy document into a beautiful, genuinely PLAIN-ENGLISH self-contained HTML explainer that a non-technical or busy stakeholder can actually understand — built on a single everyday metaphor carried through the whole piece, with the REAL numbers shown honestly beside each plain claim, an "engineer's note" layer so technical readers aren't shortchanged, an honesty/limits box, and an arcade/pixel aesthetic that opens straight from file:// (no build, no server). Use this WHENEVER the user wants to explain technical work in plain English, make a pretty/shareable report, one-pager, or recap, summarize an investigation/session "so my boss/team/client can understand it," translate jargon results for a lay audience, or asks for an explainer/writeup/"make this readable" of findings — even if they don't say "HTML." Also use to turn an existing analysis doc or markdown report into a friendly visual explainer. Prefer this over a plain markdown summary when the audience is non-expert or the user says "pretty," "plain English," "for my boss," "easy to understand," or "report."
video-prd-dual-agent
by wantedfast
Controller skill for mandatory real multi-agent Chinese video PRD production for HyperFrame / HyperFrames. Use when the user invokes video-prd-dual-agent or provides SRT, narration, screenshots, charts, product notes, statistics, or visual assets and asks for a video PRD, video generation PRD, storyboard, shot plan, animation plan, or HyperFrame-ready production brief. This skill requires visible role-bounded sub-agent dispatch for Agent A, Agent B, Scene Planner, Scene Director, Motion Agent, Text Agent, and PRD Reviewer; it must not silently simulate those roles inside one agent.
vlog-planner
by Vibetool
Plan a travel-vlog shoot from a destination + rough route + dates. Combines live weather, sunrise/sunset & golden-hour light, sun direction, and map/POI data into a day-by-day 拍摄手册 (shooting manual): for each spot it picks the best time window, scores 出片指数 (shootability), suggests A-roll/B-roll and 三镜头法 shots, and gives a weather plan B. Use when the user wants to plan what/where/when to shoot for a 旅行 Vlog / 行程拍摄规划 / 视频号·抖音·YouTube travel video — e.g. "帮我规划去稻城的 vlog 拍摄", "下周去重庆怎么拍 vlog", "plan my Iceland vlog shoot". Inspired by 博主「超Carry的柴西」's method.
claris-filemaker-pro
by CadenceUX
REFERENCE skill — FileMaker Pro script steps, calculation functions, and custom functions ONLY. Use for: syntax lookups, function parameters, usage examples, and live doc fetches across all 368 built-in functions (If, Case, Let, While, ExecuteSQL, JSON, AI/embedding, Persistent Data), all 215 script steps (including FM 26 PDF Files and Persistent Data categories), Data API (REST), SQL/ExecuteSQL, WebDirect, FileMaker Go, error codes, and custom function patterns. Not a platform administration guide. Out of scope: FileMaker Server admin, Claris Connect, Claris Studio, ODBC/JDBC deep configuration. Trigger on any FileMaker function name — even inside code, a calc field, or a script. Always check local reference files first for format, doc_url, and originated_in_version before fetching live docs. Prefer live doc fetches over training data — Claris docs are versioned and frequently updated.
Deft Framework
by deftai
Best practices and guidelines for AI agents in software development
finnskogarna
by cogpros
'Master accordion: fold N parallel Claude Code sessions when they overwhelm. Compress each, diff, wake ONE, seal or HELD the rest (HELD = quarantined-but-present, surfaced as a daily count heartbeat). Never merge. Use when (1) "finnskogarna" / "fold the sessions"; (2) N≥3 sessions pile up; (3) overwhelmed by sessions; (4) pre-bed reconcile; (5) a thread burns and should be held, not cut. NOT for single-session checkpoint (/accordion), session close (use your session-close skill), reorientation (/whats-next).'
cc-connect-wechat
by Bj2020wang
Install and configure cc-connect to connect personal WeChat (个人微信) with Claude Code as an AI assistant. This skill covers the full setup workflow on Windows: downloading cc-connect, binding WeChat via ilink protocol, configuring config.toml, registering as a Windows service with NSSM for persistent background operation, and RDP-specific fixes. Triggered by requests like "install cc-connect", "connect WeChat to Claude Code", "微信接入 Claude", "cc-connect 安装", "微信 AI 助手 setup".
schematica
by tester2024
"Schematica is a Python toolkit and interactive REPL for building Minecraft Java Edition schematics, exposed as an AI skill. Use this skill when the user asks to build, generate, design, or export Minecraft structures, schematics, .schem/.schematic/.litematic files, voxel maps, or 3D builds programmatically; when the user wants a procedural or creative toolkit for Minecraft builds; when the user mentions shapes, polygons, geometry, heightmaps, meshes, sessions, previews, Sponge schematics, MCEdit schematics, or Litematica; or when the user wants to script Minecraft map construction from Python or a REPL with undo/redo, block palettes, and multi-view PNG previews. Also use for tasks involving minecraft-data block catalogs, blockstate validation, trimesh/shapely geometry voxelization, or Perlin terrain. The AI agent is the creative driver: it composes shapes, blocks, and generators from this toolkit to realize the user's request."
super-assistant-framework
by LiubaiCheng
超级助理Agent的完整系统脚手架。包含7层架构目录结构、WIKI知识库管理规则、工程纪律方法论、自动化日程模板、setup引导流程和使用指南。当用户想要搭建一个拥有知识库管理、晨读信息摄入、信号追踪、周期巡检等自动化能力的超级助理Agent时使用此技能；当用户提到"搭助理框架""初始化超级助理""搭建知识库Agent""复制助理系统"等意图时触发。
ITOps Agent Platform 开发助手
by qinshihu
国内首个企业级 IT 运维多 Agent 自动化平台 — 基于大语言模型的智能运维解决方案。ITOps Agent Platform通过可视化工作流编排，将多个AI Agent组合成智能运维自动化流水线，实现服务器管理、告警处理、故障诊断、日志分析、脚本管理、定时运维任务的自动化执行， 支持国内外主流大模型，旨在 Zabbix/Prometheus 告警自动修复闭环，Docker 一键部署，多平台兼容。
suno-click-fix
by okihiro-ai-creative
Suno生成曲の冒頭に出るクリックノイズを、FFmpeg adeclick で問題区間だけ局所処理し、残りは原音保持＋クロスフェードで自然結合する部分修正スキル。
ai-insurance-advisor
by mnetfairy
中国大陆保险AI助手。当用户询问以下内容时使用：保险配置、保险方案、产品对比、重疾险/医疗险/寿险/意外险/储蓄险推荐、保费计算、保障缺口分析、需求分析、核保合规、理赔、朋友圈文案、培训话术、代理人展业支持。
html-explainer
by ds-vibe
Build a polished, lightly-interactive single-page web explainer for educational, informational, or explanatory content. Use when the user wants to explain, teach, summarize, or visualize a topic, concept, policy, dataset, process, or body of knowledge as an interactive web page — emphasizing progressive complexity, strong visual design, source-grounded accuracy, and an iterative quality loop. Triggers on requests like "make an interactive explainer / guide / walkthrough / one-pager," "turn this into a visual web page people can explore," or "help people get acquainted with X and drill down."
e2e-test-skill
by Claire528717
"Design PRD-driven end-to-end acceptance test plans from product requirements and real UI sources. Use when Codex needs to turn a PRD, screenshots, prototype, HTML, frontend code, or other visual UI source into executable E2E test cases with realistic UI paths, mock/test data plans, review gates, and independent database/API/external-system verification. Also use for Chinese-language requests such as 验收测试, 端到端测试, E2E 用例, 测试计划, or 根据页面生成测试用例 after development and integration are complete. Execute cases, collect evidence, handle bugs, and produce run summaries only after the user approves the plan and explicitly asks to run it."
editing-skill
by LuoJiangYong
"When the user asks to edit AI/tech news, write daily/weekly/monthly reports, analyze topics, recommend GitHub projects or products, generate social insights, or create WeChat HTML articles — use this skill for structured news editing and format consumption."
temis-dmn-engine
by pblumer
Betrieb und Deployment von temis — der Go-basierten DMN 1.5 Decision Engine mit FEEL-Support. Build from source, Deploy via Ansible, Model-Management, MCP-Nutzung.
l-route-reflection-skill
by sushuqiong
"Turn unstable-platform career narratives into privacy-safe reflective outputs, including timelines, essays, comparison tables, image prompts, and route reconstruction."