Browse agent skills

Declarative provider, model, connection, and secret composition for Clear Ideas Agent Runtime.

Streaming prompt and ordered tool-call executor for Clear Ideas Agent Runtime.

Persistence-independent deterministic execution core for Clear Ideas Agent Runtime.

WordPrint — AI content detector CLI + web UI (Grok, Claude, or OpenAI). Scores how LLM-generated a passage looks.

n8n community node for DeepSeek V4 models (deepseek-v4-flash, deepseek-v4-pro) with thinking mode control

Installable TradingAgents multi-agent stock-analysis skill (Claude Code / Cursor / Windsurf / Trae / OpenCode via vercel-labs/skills)

Interactive TUI for discovering and managing local AI model endpoints (llama.cpp, oMLX, Ollama, vLLM, SGLang, LM Studio). Reads real server-reported model configurations — no hardcoded databases.

Detects hidden text and prompt injection in PDF and DOCX documents: white-on-white text, sub-legible fonts, off-page text, invisible render mode and invisible Unicode characters.

Zero-dependency EMVCo QRPh toolkit — decode, encode, and generate GCash-compatible QR codes with pre-filled amounts. Works standalone or as a Discord bot plugin.

Shared contract layer for the pi-* packages: one taggedDigest, one project identity, one set of event payloads with make*/parse* pairs, and the knowledge-signal/replay-port types. Zero runtime dependencies; ships compiled dist only.

AI Cost Guard n8n community node generated by NEXUS Plugin Factory.

Generate marketplace listing copy, buyer setup docs, support boundaries, and pricing notes for n8n workflow templates.

Official TestNeo community node — agent-run ingest, post-agent verification (PASS/BLOCK), and release outcome recording. Same REST contracts as the TestNeo n8n HTTP templates.

Stable terminal UI for Codex and Claude Agent SDK

Playwright story gallery for component testing

Lean output mode for Claude Code: minimal solution, action-first structure, zero-fat density. Fuses caveman, ponytail, and i-have-adhd under one rule.

`react-edit-to-html` is a premium, highly customizable, responsive HTML editor and visual template builder for creating dynamic HTML layouts. It is perfect for generating backend templates (invoices, shipping lists, email templates, billing letters, repor

Convert provider OCR responses into openparser@1 documents

Standalone CPU and GPU inference runtime for Uni-pKa

TypeScript SDK for the SSP1126 (Saman Electronic Payment) POS terminal — ISO-8583 over TCP/Serial

VMI LLM SDK — prompts, memory (vector + connected DB), transactions, multi-provider LLM calls

ReAct Agent framework for coding tasks

In-memory full-text search engine for Bun and Node, implemented in Rust; MiniSearch-compatible

A production-grade, strongly typed Node.js SDK for the Higgsfield AI API.

Coding agent CLI with read, bash, edit, write tools and session management

Thin, drop-in Python SDK for the 1xAi gateway (https://1xai.ir) — OpenAI, Claude, Gemini and DeepSeek from inside Iran, billed in Toman.

AI subscription usage gauge for the pi coding agent — Claude Pro/Max and ChatGPT Plus/Pro quota bars with provider status (incident) alerts, in the footer. Model-aware fork.

Agent skill for art-directed, CMS-native, Lighthouse-verified websites: source-world art direction, screenshot critique loop, and a scanner for AI-template design

n8n community node for Auphonic: a trigger that starts a workflow when a production finishes processing, and an action to create and start new productions.

Semantic diff review for coding agents: an LLM turns a git diff into a narrative HTML report, a human comments on it, and the feedback comes back as plaintext.