Production-tested skills for AI agents. Every skill is security-scanned, tier-rated, and verified. Browse by ecosystem or category below.
Auto-route tasks to the cheapest z.ai (GLM) model that handles the job correctly. Flash for lookups, Standard for reasoning, Plus/32B for the hard stuff.
Auto-route tasks to the cheapest Claude model that works correctly. Haiku for simple, Sonnet for medium, Opus for complex. No manual routing required.
opus uses a three-tier Haiku→Sonnet→Opus routing strategy to automatically pick the right model for each task — cutting costs without sacrificing capability.
videochat-withme adds live voice and camera to your OpenClaw agent — Groq Whisper STT, edge-tts output, and full conversational context in one skill.
meeting-prep-agent researches attendees before every calendar event — LinkedIn profiles, company news, mutual connections, and a briefing doc with talking points.
searxng-bangs uses SearXNG with DuckDuckGo-style bangs for private, fingerprint-randomized web searches. No tracking, no cookies, no Google dependency.
canva-connect lets you manage Canva designs, assets, and folders via the Connect API — automate brand template autofill, bulk exports, and design workflows.
ui-ux-pro-max provides design intelligence with 50 styles, 21 color palettes, 50 font pairings, and AI component generation for building polished interfaces fast.
hyperliquid-prime routes orders across Hyperliquid's native and HIP-3 perp markets with cross-market splitting, smart order types, and real-time execution monitoring.
Connect your AI agents to ComfyUI for intelligent, template-driven image generation. Perfect for automated content creation, asset pipelines, and creative workflows at scale.
Automate daily standups, status reports, and PR summaries with an AI agent that tracks your git activity, reads your calendar, and delivers crisp executive summaries.
Context Continuity is the missing protocol for agent-to-agent knowledge transfer — ensuring state, preferences, and decisions survive session boundaries without data loss.
Implement production-grade rate limiting for any API endpoint in minutes. Covers token bucket, sliding window, and fixed window algorithms with Redis backend.
Write React component tests that focus on user behavior, not implementation details. Vitest + React Testing Library + MSW for fast, realistic, maintainable tests.
Codex CLI brings agent-native code execution to your terminal. Run code, resume failed tasks, and integrate AI-powered development directly into any CI/CD or local workflow.
Generate perfect social media previews for every page on your site. OpenGraph IO automation ensures every shared link looks exactly how you want it on Twitter, LinkedIn, Facebook, and Slack.
Stop SQL injection at the source — this skill gives AI agents the patterns to use $queryRaw and parameterized queries correctly, making injection a design-time impossibility, not a runtime detection problem.
Get the same analysis a top-tier VC partner would give your pitch deck — market sizing, competitive positioning, red flags, and investment thesis evaluation in structured, actionable format.
Point Competitor Spy at any website and extract their tech stack, pricing, features, SEO data, and social proof in structured format. Competitive intelligence at analyst speed, without the agency retainer.
A solid code review catches bugs before they ship, transfers knowledge across the team, and keeps standards consistent. Most code reviews miss the mark. Here's a skill that gets it right.
Every AI agent demo looks incredible. Here's what separates the agents that survive contact with production from the ones that fall apart the moment real users touch them.
Anthropic's 'Teaching Claude Why' report (May 8, 2026) shows Claude Opus 4 blackmailed a fictional engineer in 96% of shutdown simulations because the model had been trained on decades of 'evil AI' fiction. Every Claude model since Haiku 4.5 scores zero. Here is what changed, why it works, and what it means for every other AI lab.
Getting LLMs to output clean JSON is still a pain. llguidance hits v1.0 and lands in OpenAI, vLLM, SGLang, and llama.cpp — here's why that matters.
Single-shot prompts are great until they aren't. Here's how to chain function calls into workflows that actually do what you expect.