Turn any screenshot/design reference into a precise, copy-paste-ready developer prompt. Upload a screenshot → get exact hex colors, Tailwind classes, animation specs, and component structure — ready for Lovable, Bolt, Replit, or Cursor.
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 24,091–24,120 of 127,302 skills
Page 804 of 4,244 · loading 30 at a time
Lint your linting: audit a repo's API-governance CI setup against an 8-point maturity rubric and get a prioritized punch-list of fixes. The sequel to spectral-reporter. An API Commons tool.
The shared Playwright capture + flow engine for the cardmem-lens fleet — capture(opts)→artifact and runFlow(opts)→report with self-healing locators (DOM layers + Set-of-Marks vision) and a frozen Zod flow grammar, plus token-frugal page-READ primitives (r
AI Lane — inspect your hardware and find out which local AI models it can run
A Minecraft-themed statusline for Claude Code — hearts, food, and XP bars for your rate limits and context window
Fayz SDK — reusable auth runtime and UI integration plugin
Fast WebGPU runtime for 1-bit (binary-weight) LLMs in the browser. Bit-exact, zero runtime dependencies.
Claude Code skill that teaches you anything in one screen: visual lessons with memory and a feedback loop
Catch the LLM/AI calls you didn't need. A fast, deterministic linter that flags LLM API calls where plain code is simpler, cheaper, and more reliable.
Self-contained PKCS#7 PDF signing — no SaaS, no metering, no per-doc fees. Render→sign→verify engine + pluggable persistence interfaces + end-to-end signDocument() orchestrator.
OpenScience is a model-agnostic, open-source AI research agent for scientific and ML engineering work. It runs a workspace in your browser where the agent plans tasks, writes and runs code, drives experiments, queries scientific databases, and writes up r
Vercel AI SDK tools for the Arcis treasury — earn yield on idle USDC on Base mainnet. Works with any model via the AI SDK.
One OpenAI-compatible endpoint for 15+ free LLM providers. Smart routing, failover, encrypted keys. Run with npx freellm.
Atomic, GitHub-native mutex for multi-agent work. The lock IS a git ref — no server, no database. First writer wins.
Let Claude (or any LLM) actually watch a video — scene-aware, deduplicated frames + transcript, from a URL or local file.
Bash policy enforcement — provides a bash_readonly tool that blocks mutating commands
Nexrall Code — AI coding assistant for your terminal (headless agent for scripts, CI and automation)
Shared parser, schema, validation, and install target logic for PapiSkill packages.
Vercel AI SDK tools for AudioLab — loudness, voice-quality, and signal analysis via the AudioLab hosted API.
An AI-first game engine: a deterministic, headless-native simulation kernel with a Godot-style scene tree, pluggable renderers (SVG/Canvas/headless), and a built-in verification harness — designed so an LLM can author, test, and prove correct a whole game
code — local RAG coding assistant powered by Ollama and LanceDB
Production-ready n8n community node for the official eSewa ePay API
CLI to install personal Claude Code skills
Cloudflare Workers AI Account ID & Token Auto-Grabber — login OR signup from scratch
CLI tool and web client to transform Ecuadorian electronic invoice Excel files from 26 columns to 12 columns, with dynamic TIPO GASTO selection, custom output filenames, and a modern web interface.
OMP Studio agent (omps): open, sync and version Grenton .omp projects in the hosted editor