Browse agent skills

GraphQL operations as @nxgt/ai-core tools: an allow-list run in process with graphql-js, forbidden fields rejected at build, mutations behind a confirmation

alxia glue for @nxgt/ai-core: a request middleware (abort signal, language, observers) and a Server-Sent Events reply that streams an agent run

Pyramid Principle skill for Claude, in Spanish and English · Skill del Principio de la Pirámide de Barbara Minto para Claude

pi extension that sends calibration plots to a calibration-evaluation VLM such as NVIDIA Ising-Calibration and returns a structured diagnosis

Connects Coo to a QQ bot entry: QR binding, direct and group messages, stickers and media, with QQ's reply limits visible to the model

Server management, an independent encrypted credential vault, and per-server recurring operations for DeepSeek Harness.

SigTerm behaviour classifier: terminal windowing, the rules baseline, the open-jev inference interface and its tunable labels file.

pdfplumber for JavaScript: extract chars, words, text, lines and tables from PDFs in the browser and Node, with the same results as Python pdfplumber.

Git remembers what changed. cogit remembers why.

One door to every model: a Model takes text, image, video, audio and PDF parts, answers with text, tokens and USD, and parses an answer to a zod shape. Adapters for Anthropic, Gemini, Cohere, any OpenAI-compatible server and a shell command (claude -p); e

Apple Intelligence for Expo and React Native: on-device Apple Vision, Foundation Models (the Apple Intelligence LLM), SpeechAnalyzer and Image Playground, from one decoded image.

쓰던 앱의 기능과 흐름을 내 앱으로 다시 짓는 무료 Claude Code 스킬 11개 묶음(replica-skill)을 공식 방법 그대로 한국어 안내 스킬로 넣어 줍니다. 스킬 본체는 제작자의 것이며 싣거나 대신 설치하지 않습니다.

Convergent deliberation loop for coding agents: multi-pass, coverage-gated, append-only ledger review and planning. Claude Code, Codex, OpenClaw.

Playwright reporter that keeps a QA dashboard up to date after every run: health score, trends, flaky tests, failure clusters, Jira bugs, Notion coverage. Open it when you need it.

Documentation compiler and CLI with HTML, Markdown, and read-only MCP tools.

The aura command line: service keys and agents built from a spec, for people and coding agents

LLM benchmarks & evaluation tools — measure what matters

React bindings for poly: provider, hook and a ready-made prompt that offers to switch the interface to the language people speak.

Voice-driven interface translation: detects the spoken language and translates your UI texts on the fly, with protected terms, reviewed overrides and caching.

Local n8n node to send Telegram messages through a pinned IPv4 address while preserving TLS hostname verification.

An @nxgt/ai-core agent observer that writes runs, model calls and tool calls as @nxgt/telemetry spans with gen_ai.* attributes

Ollama /api/chat adapter for @nxgt/ai-core on Bun fetch, with no third-party client

JavaScript client for the Comet REST API, and the base classes for Comet JavaScript panels

Reactive LLM eval, tracing, and regression layer as a Convex component.

An open protocol and Agent Skill for measuring how well your app or game resists AI-assisted reverse engineering. Works with Claude Code, Codex, Gemini CLI, Cursor and more. Scan your build, get a resistance score, see what leaks, and learn how to fix it.

n8n node for JustGains: log workouts, food and bodyweight, and read your training history, progress and insights

A way of working that Product, Developers, QA and AI share: plain-English briefs, specs the AI translates rather than authors, and debt it has to declare.

A convenient Playwright wrapper for browser automation. One line to open a browser, then navigate, click, fill forms, scrape links, take screenshots and collect downloads — with full access to Playwright underneath. Same API as puppet-browser-handler.

Prompt seguro de senha do sudo para o Pi Coding Agent — modal mascarado via SUDO_ASKPASS; a senha nunca passa pelo modelo

Signal House SDK for use with the Signal House platform