Loading...
Directory
Search by intent: design systems, testing, deployment, security, documentation, Azure, Supabase, React, or any workflow you want an agent to perform better.
Showing 1,351–1,380 of 116,534 skills
Page 46 of 3,885 · loading 30 at a time
Real-time local speech-to-text for the DeepSeek Harness composer: live partials while you speak, SenseVoice finals with punctuation, everything on-device · 实时本地语音转文字：边说边出字，松手写入输入框
launchBrowser(ctx) for Offsida browser steps: Playwright through the platform's Browser Run gateway, or a local Chromium under the same policy in tests.
Temporary package placeholder for staged publishing
Terminal coding agent with skills, MCP, memory, hooks and free-tier model routing
A unified API for calling multiple LLM providers through a consistent, OpenAI-compatible interface
Steering vectors as a training signal in LLM post-training (NeurIPS 2026)
Interactive editing overlays for rasam SVG drawings.
Shared ESLint flat config: JavaScript and TypeScript recommended, SonarJS, React (@eslint-react), React Refresh, plus Vitest, Testing Library and Playwright presets
Reelsmith: agent skill that turns a voiceover, scene clips, a script and a logo into a finished Instagram Reel. Local Whisper captions, HyperFrames edit, moving logo watermark, 30 fps. Claude Code, Cursor, Codex, OpenCode, Antigravity. EN/TR.
The domain core: canonicalisation, Ed25519, `did:key`, schema validation, and mandate build/sign/verify/hash, plus the shared action vocabulary. A mandate records who lets which agent do what, within which limits, until when, for any action in any busines
Turn reusable workflows from finished DeepSeek Harness sessions into Agent Skills after review
The server SDK for Flatstack apps: functions, routes, schema, storage, cron, realtime and sign-in settings.
Free Claude skill, built for Kie.ai: turn an ad script into a finished video ad from AI clips, your own clip library, or both.
Stand up an OpenAI-compatible LLM endpoint on a rented or owned GPU box: pick the instance, fetch and quantize weights, serve with SGLang — speculative decoding, prefix caching and tool calling for agent workloads — then verify it and expose the port.
Agent skill that makes Claude Code, Codex, and Cursor write and review backend code like the engineer who gets paged at 3am. 90 reliability and security rules.
Evaluate decision models (Jev, Clef, d1, OpenAI Decisions) on your own data, then run the one you trust in your app and CI
Portable native RapidOCR NCNN, with Chinese/English models; no Python
Temporary package placeholder for staged publishing
Personal knowledge wiki (second brain) with Claude Code on the nvk/llm-wiki plugin: method, installers, scripts and AI specialists. No data, only the method. English, Russian and Ukrainian.
AXI-compliant Argo Workflows CLI wrapper - token-efficient TOON output, read-first, previewed and allowlisted writes
Vue bindings for @asrjs/speech-recognition: in-browser speech recognition state in your components.
Agent-ergonomic Argo CD CLI with opt-in writes - AXI-compliant argocd wrapper with token-efficient TOON output
Svelte bindings for @asrjs/speech-recognition: in-browser speech recognition state in your components.
React bindings for @asrjs/speech-recognition: in-browser speech recognition state in your components.
Speech-first TypeScript runtime for browser and local Node.js inference
Zero-dependency Bring Your Own Key (BYOK) AI toolkit for Browser, Desktop, Mobile, and Node.js CLI applications.
Publication-ready matplotlib figures for academic papers, as a Claude skill. IEEE/ACM/USENIX/ICML/NeurIPS presets, 9 chart forms, colorblind-checked palette.
Agent skill: a general testing discipline — a requirement is covered only when a test shows that violating it fails. Business rules, data invariants, persistence, concurrency, API contracts, security.
Independent vectors, matrices, quaternions, numerics, statistics, units and control math