Split streaming LLM output into sentences or Markdown blocks as soon as each one is complete. For text-to-speech, live translation and chunked rendering. Handles abbreviations, decimals, code fences, lists and CJK. Zero dependencies.
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 3,061–3,090 of 125,967 skills
Page 103 of 4,199 · loading 30 at a time
Cortico World: 音频理解(Qwen2-Audio 本地模型)——听一段音频,说出它在放什么(音乐类型/情绪/歌词大意),也能把语音转成文字。
Dispute and refund verdicts for AI-agent payments: Predge tools for the Vercel AI SDK, plus signed Polymarket settlement-risk checks
Voicely text-to-speech integration for n8n workflows
Training Lens · 训练视窗 — a local, read-only TensorBoard scalar dashboard for DeepSeek Harness.
把文章、案例或笔记排成 3:4 小红书图文 | 1080×1440 | 20 个母版 | 浅色 / 深色 | Claude Code Skill
Charts from a Markdown table, drawn to SVG while parsing, that copy as a PNG or as data: bars, stacked bars, lines, areas, scatter plots, dumbbells, waterfalls, pies, gantt charts, and sankeys. The table stays in the page, so people see the chart and agen
Zero-dependency streaming CSV parser and writer for Node.js: RFC 4180, GB-size files with flat memory, schema validation, safe type casting, formula-injection protection, encodings and precise errors.
Claude Code skill: a startup co-founder that asks one question at a time and builds your Startup Plan (problem, ICP, v0/v1/v10, TAM, GTM, pricing, fundraising)
Analyze and transform faces in n8n through the Ace Data Cloud Face Transform service.
Free IDE skill that turns a GitHub pull request into a two-zone review map. Use it in VS Code, Cursor, Claude Code, or Windsurf. Your model reads the PR; the map opens at pr-viz.com or on your own host.
Vendored skill-merge Node pipeline engine (byte-identical upstream) for DSH evolution-merge (community build)
Python SDK cho Locaith-4-flash: tìm web, suy luận, ngữ cảnh 64K và Bio-Memory.
memcue long-term memory for DeepSeek Harness (dsh): the model searches and writes its facts itself
Official Python SDK for the Yunmail cloud communications API — email, SMS, voice verification and push notifications through one unified API.
Import .jsonc files, JSON with comments, as typed modules, read with jsonc-parser.
Import .csv and .tsv files as typed arrays of rows, read with csv-parse.
AI brain viewer for DeepSeek Harness: a brain-shaped relationship graph in the left sidebar, plus a Markdown knowledge-base pipeline (import -> compile -> graph).
Whisper for pipedog: the base service for Whisper-compatible APIs (shared by the OpenAI and Groq STT services), probability helpers, and local Whisper models (WhisperSTTService) on ONNX Runtime with faster-whisper's decoding.
WebSocket transports for pipedog: client, single-client server and bring-your-own-socket (FastAPI) transports over a runtime-agnostic socket, with servers for Bun, Node (ws) and Deno.
Unified Scrapeless CLI for Web Unlocker, Crawl, AI Scraper, proxies and cloud browser automation
DeepSeek Harness plugin that renders Inception Mercury's diffusion (denoising) stream live in the Web UI
Directus flow operation to send Slack messages and Block Kit layouts through a webhook or bot, for notifications and alerts
GitHub Actions reporter for Playwright
AI provider runtime for Gryphon: one interface over the Claude Code, Codex and Antigravity command-line assistants and the Anthropic, OpenAI and Google APIs, with tool loops, streaming, cancellation and retries.
NexLLM image provider for the Vercel AI SDK (OpenAI-compatible POST /v1/images/generations)
DSH skill distilled from the whole of Gillespie's Fundamentals of Vehicle Dynamics (10 chapters, 164 capability cards): axle loads, acceleration, braking, road loads, ride, steady-state steering, suspension geometry, steering systems, rollover, and tyres
Unofficial Tistory CLI: publish Markdown posts from your terminal via browser automation