Browse agent skills

Voice input and output for pi powered by Xiaomi MiMo V2.5 TTS

Long-term spoken-English coaching tracker for Claude Code: practice by voice in the Claude app, get scored reports, a mistake bank, trend charts, and an adaptive next-session focus.

Static security & correctness linter for Payload CMS (the headless CMS) — audits collections, access control, hooks, routes and config. Not an API-payload validator.

Usage and cost visibility for Claude Code and Codex. Cross-platform terminal CLI (macOS/Linux/Windows) with ccusage-style reports + a live 5h-block dashboard.

Official JavaScript/TypeScript SDK for the Sentinel pay-per-use AI API marketplace on Algorand

Drop-in floating AI translation button for any React app — runs 100% in-browser, no API key needed.

Official npm launcher for Siyarix — AI Cybersecurity Orchestration Agent

AI-powered TextFSM template generator with multi-provider routing

Metadata-first AI framework — the LLM receives only schema and column names, generates deterministic code, executes locally on your data

Databricks Model Serving and Vector Search tools for Chidori TypeScript agents.

AI Coding Skeleton — CLI + LLM协作架构: CLI搭建骨架 → AI深度初始化

Zero-config LLM API inspector — intercepts calls, shows token breakdown and cost estimates in a local dashboard

Visual Retrieval-Augmented Generation — render, embed, index, search, train

Gatebreaker: A brutally honest, god-tier cybersecurity career diagnostic system prompt, agentic skill, and simulation dashboard.

Client SDK for mikser-io's vector plugin — semantic search from the browser or Node

The Memory Layer for Enterprise AI — one API swap adds persistent memory and 40% token cost reduction to any LLM stack

Unified LLM API with automatic model discovery and provider configuration

Sparkling port of expo-speech (scaffold — name reserved for future implementation)

Cross-platform installer that deploys a dual-system (Claude Code + opencode) AI-coding engineering rule system: global constitution, auto-triggering skills, and enforcement hooks.

Open-source code + knowledge graph memory engine (SAGE + CodeGraph). AsiaOstrich is only a reference consumer.

🍩 Donut Corp — local LLM CLI, API & desktop app powered by Donut Runtime.

Elite frontend design taste for Claude Code / Cowork — a merged synthesis of emilkowalski/skill, pbakaus/impeccable, and leonxlnx/taste-skill. Typography, color, motion, components, anti-slop.

Save ChatGPT conversations and JavaScript-rendered pages to searchable PDF or plain text

Local session transcript memory for Claude Code and Codex — search past sessions with FTS5 + semantic vectors.

Coolify CLI agent skill — teach your coding agent to deploy apps, manage databases/services, and stream logs on Coolify Cloud or self-hosted.

Generic finite-state-machine kernel for multi-agent LLM workflows — atomic state, replay-deterministic, zero runtime dependencies.

DeepSeek prefix-cache extension for pi — hit-rate telemetry, prefix guard, and cache-friendly compaction.

Command-line tool and API to extract Microsoft Excel data.

Framework adapters for mosadd. Subpath exports: @mosadd/ai/vercel, @mosadd/ai/langchain, @mosadd/ai/openai, @mosadd/ai/anthropic. Single package, atomic releases (Stripe Agent Toolkit pattern).

🔍 Claude Code Skill: Catch the #1 AI anti-pattern before you build. Reviews tool designs with a 4-point decision tree — AI-decorated vs AI-native.