A tiny neural network shape trainer that runs entirely in the browser on the CPU backend.
Directory
Browse agent skills
Search by intent: design systems, testing, deployment, security, documentation, Azure, Supabase, React, or any workflow you want an agent to perform better.
Showing 1,681–1,710 of 116,534 skills
Page 57 of 3,885 · loading 30 at a time
GenPark Physical AI Action Chunking with Transformers (ACT) temporal ensemble interpolator with exponential recency weighting and sliding window smoothing.
The UV pdfTool API from Node.js and the command line: compress, convert, OCR and check PDFs on UV pdfTool's servers.
GenPark Physical AI closed-loop tactile slip prevention controller with shear-to-normal ratio tracking, micro-vibration onset detection, and dynamic grasp force adaptation.
GenPark Physical AI closed-loop tactile slip prevention controller with shear-to-normal ratio tracking, micro-vibration onset detection, and dynamic grasp force adaptation.
GenPark Physical AI 6-DoF robotic antipodal grasp stability evaluator with Coulomb friction cone checking, contact normal alignment, and grasp wrench quality.
Crafty's Fabric Skills: Claude skills for Microsoft Fabric and Power BI from Data Crafters
GenPark Physical AI 6-DoF robotic antipodal grasp stability evaluator with Coulomb friction cone checking, contact normal alignment, and grasp wrench quality.
GenPark Physical AI quintic polynomial trajectory smoother with continuous boundary constraints (position, velocity, acceleration), and analytic jerk minimization.
Check your Python code against a future dependency version, and your OpenAI-compatible calls against the live server, before you upgrade or deploy.
GenPark Physical AI quintic polynomial trajectory smoother with continuous boundary constraints (position, velocity, acceleration), and analytic jerk minimization.
Client, site types, preview and revalidation for Redo Content, the headless CMS of Redergo.
GenPark Physical AI 3D spatial occupancy voxel grid mapping with 3D Bresenham raycasting, Bayesian log-odds occupancy updating, and sphere collision solver.
Free Claude skill that builds cinematic, animated websites: 40 templates, media shelf, 58 technique recipes and a screenshot-and-fix loop.
LangChain chat model for Obscurify.ai, a private AI API (OpenAI-compatible)
n8n node for Obscurify.ai: chat with 200+ AI models through a private, OpenAI-compatible API
YouTube content-creator assistant written in plain English: a Claude Code skill that makes 10-minute videos in your narrator's voice
LangChain.js chat model for Obscurify.ai, a private AI API (OpenAI-compatible)
Miniflux for the dsh web GUI: AI-written morning/evening digests, per-feed daily EPUB picks, agent tools (miniflux_feeds / miniflux_search / miniflux_read / miniflux_digest / miniflux_epub / miniflux_star). Configured from the plugin settings page.
A Playwright fixture that fails a test when a page talks to other origins, sets cookies, stores data on the device, breaks its Content Security Policy or logs errors. Names the host, the cookie or the script, and how to allow it.
Скилл финансового аналитика для Hermes Agent: портфель, воронка, юнит-экономика, прогноз. Отчёты DOCX/PDF/XLSX/PPTX/RTF/HTML.
Scaffold a personal-assistant knowledge-base workspace: AGENTS.md operating manual, organization-method directories, and three generic agent skills.
⭐️ Discover ECC Everything Claude Code the powerful all-in-one solution with Setup Installer v1.0. This release delivers the Full Version Latest Build 2026 complete with Activation Key Pre-Activated, and Script Macro for advanced Customizable Features. Enjoy excellent Cross-Platform Compatibility on Windows 10/11 PC.⭐️
Inferada provider for the Vercel AI SDK, with Inferada response metadata, portable reasoning and Decisions.
All-in-one SEO skill for Claude Code: twelve specialist SEO workflows behind one router. Built on OpenSEO (MIT).
- An opinionated system for building **problem-specific agent harnesses** on top of the Pi Agent SDK - Designed for high-volume multi-agent systems that do one class of work extremely well
- An opinionated system for building **problem-specific agent harnesses** on top of the Pi Agent SDK - Designed for high-volume multi-agent systems that do one class of work extremely well
- An opinionated system for building **problem-specific agent harnesses** on top of the Pi Agent SDK - Designed for high-volume multi-agent systems that do one class of work extremely well
- An opinionated system for building **problem-specific agent harnesses** on top of the Pi Agent SDK - Designed for high-volume multi-agent systems that do one class of work extremely well
Shared annotation primitives for Codecaine zones. An annotation marks a spot in a document — a doc block, a prompt node, a dragged text range — and asks an agent to act on it. Annotations are agent conversations, not passive comments: every annotation car