Browse agent skills

Node.js/TypeScript client SDK for third-party partners to call the Jyotish interpretation gRPC API.

Trajectory state entropy and cyclic repetition detector identifying autonomous agent infinite tool loops and planning deadlocks.

bot-telekit is a modern, type-safe Telegram Bot API library for TypeScript

Claude Code skill: brief + references → beat-synced single-file canvas motion graphic (HTML) with synthesized sound, mobile player and automated verification.

Guardrails for the agents you build: permissions, verification, approvals, grounding and budgets from Scute, on the agent framework you already use.

Convert a SVG to multiple image formats (without puppeteer or a headless browser)

Next-level LLM code review: version-aware skills, static hints and self-critique for fewer false positives. Works with Claude Code, Codex, Copilot, Gemini, the Anthropic API, AWS Bedrock and OpenAI-compatible APIs (Ollama, OpenRouter, …).

Typed, probabilistic decisions with Jev by TypeSafe AI for n8n workflows

n8n community nodes for the complete DealerAI API

Lokale Dateisystem-Nodes für n8n

n8n community node for TypeSafe System One (Jev): ask Noul, Choice and Score questions about any item, branch on calibrated probabilities, and send low-confidence answers to human review.

Host-independent audio recording, input device discovery, and Whisper transcription

Speak AI-generated audio description in the gaps between dialogue, on the TV itself: duck the film, play the cue, restore — driven from the remote, with three verbosity levels.

O1-Code - AI-powered coding assistant

Transports for Bandwise runs: System One models through TypeSafe, OpenRouter or the Vercel AI Gateway with your own key, or recorded fixtures with no network.

Local project knowledge for AI assistants through MCP.

Rime streaming text-to-speech SDK for Node.js

A real WhatsApp Cloud API integration: webhook parsing, message rendering, delivery retries, and the 24-hour session-window rule, done correctly.

TypeScript client for Volter Harness's harness.v1 session and runtime service

Customer skill bundle for Catalyst Cloud, named from the executive's seat — get set up, see what is happening, see what needs you, run a project, unstick what is stuck — plus the catalyst CLI that reads through the Catalyst Cloud SDK, writes through the a

Instructional diagrams of KNX installations, with a limited simulation layer.

The kernel of Volter World: one log per twin, branches as pointers, checkpoints, the fold that keeps a twin current, the head that performs a write against the vendor, references, and the git plane. A twin package builds on it; the runtime serves it.

Three skills for realtime 3D on the web: the three.js WebGPU runtime with TSL, compute, post-processing and a WebGL2 fallback; the glTF asset pipeline from source to frame, with compression, budgets, sourcing and optional asset-service ordering; and a 3D

Thin LangChain integration for Komplex AI — a hallucination detector for LLM output (adapts the halu SDK to LCEL).

Sudarshana — Premium Futuristic Spiritual AI Assistant inspired by a cosmic universe

Token Larper — Multi-Harness AI Coding Usage & Cost Telemetry Dashboard powered by ccusage

Roll back, branch, and export any DSH conversation — from the UI or straight from the agent via a message_ops tool. Non-destructive: logs stay append-only and recoverable. 消息回滚/删除/分支/导出四合一，内置 message_ops agent 工具；非破坏语义，日志只追加可恢复。

Route each coding-agent turn to the cheapest model that can finish it. Switchyard-style execution signals, TypeSafe's jev as the judge, and honest cost accounting. Works with Claude Code, Codex, OpenCode, and Pi.

Cut LLM costs: route routine tasks to cheap subagent models (DeepSeek Harness plugin)

Run the Gloss server on your computer with your own OpenAI API key.