Browse agent skills

Multimodal perception for Pi — image/audio/video/document understanding, transcription & image generation via Gemini (gemini_api or local antigravity CLI).

DeepSeek Harness plugin: multimodal tools (image/audio/video/document understanding, transcription, image generation) via Gemini API or the local Antigravity CLI.

Review Insights Tool analyzes online customer reviews and turns unstructured feedback into useful business insights.

Rooyai Image / Vision Model node for n8n - Multimodal AI image analysis, OCR, visual reasoning, and question answering

Install and run Zero Voice Bridge from one command

Vinci AI — self-hosted AI shorts generator: Gemini-powered script/voice/title generation, TikTok + Pinterest video search, ffmpeg rendering, and YouTube upload/scheduler with Indian-timezone (IST) Autopilot.

An Earendil Pi extension that explains pasted text, assistant responses, local documents, public webpages, and recent session turns (as shapes) in a context-isolated window.

Credential-free concurrent web search fan-out for DeepSeek Harness (Startpage, DuckDuckGo, Ecosia, Google, Mojeek, consensus-merged, soft/hard deadlines)

Umbrella package for OpenPromo first-party Chat SDK adapters

First-party Chat SDK adapter for TikTok Business comments, built on @openpromo/tiktok webhook schemas and the comments client

First-party Chat SDK adapter for Facebook Messenger DMs, built on @openpromo/meta webhook schemas and the Facebook Page messaging client

First-party Chat SDK adapter for Instagram DMs, built on @openpromo/meta webhook schemas and the Instagram messaging client

Platform-neutral shared machinery for OpenPromo first-party Chat SDK adapters — thread identity, webhook ingress, normalization, and capability tables

Vision bridge for text-only models (DeepSeek and any route without image input): a pluggable vision model (Gemini, Groq, or any OpenAI-compatible endpoint) describes attached images at the llm/stream seam and provides the describe_image tool.

On-device AI with cloud fallback for Capacitor apps. Apple Foundation Models on iOS, ML Kit GenAI (Gemini Nano) on Android, an OpenAI-compatible endpoint as fallback — one API, with streaming and cancellation.

Eyes for text-only DeepSeek on DeepSeek Harness: image understanding via Doubao Web by default (zero cost, no API key), Antigravity IDE quota (flash/pro), Gemini API, or Cockpit proxy — model-invokable vision tool, wrapper adapters, evidence memory, and a

AI4Scholar for DeepSeek Harness (dsh): 38 native academic tools — Semantic Scholar, PubMed, Google Scholar, arXiv, bioRxiv/medRxiv, DOI resolution, PDF full text, auto-cite, scientific figures, credit balance — with an API-key card in the Web UI settings.

Google Lens API for n8n: reverse image search from any image URL. Get visually similar results, shoppable product matches with prices, or every page carrying the exact image, as structured items. Apify-backed community node, pay-per-result, usable as an A

AI image generation plugin for the dsh web GUI: text-to-image and image-to-image through configurable provider channels (gpt-image-2 / grok-imagine-image / nanobanana series / seedream-5.0-pro / dall-e-3, with native xAI Grok Imagine, Google Nano Banana a

Eyes for text-only DeepSeek on DeepSeek Harness: image understanding via Antigravity IDE quota (default, flash/pro), OpenAI-compatible VLM endpoints, Gemini API, or local Ollama — model-invokable vision tool, wrapper adapters for deepseek/opencode-go, evi

Free Google search provider for the DeepSeek Harness web seam (ctx.web): drives local Chrome through the agent-browser CLI (headless, --profile) with no paid search API

OpenCode plugin: describe images using free Gemini, Groq, and Cerebras vision models with configurable fallback order.

Vertex AI Gemini grounded web search, research, and direct URL extraction tools for Pi, using Google Cloud Agent Platform credentials.

Google Antigravity (agy) OAuth auth + model access plugin for DeepSeek Harness: multi-account pool, 429 rotation, device fingerprinting, CLI and web login.

给 DeepSeek Harness 加视觉能力：可视化设置页里选视觉厂商（Kimi/智谱/千问/OpenAI/Gemini/豆包）并填 API Key，保存后即可在聊天里选择对应视觉模型——图片交给视觉模型识别、文字交给 DeepSeek 推理。

DSH Web 语音输入插件：零服务端、零密钥、零模型下载，直接使用浏览器内置 Web Speech API（Edge=微软 Azure 语音，Chrome=Google 语音）。

Generate style-consistent icons with Google's Gemini image API, the same way a Gemini "Gem" does — system instructions plus reference images in, a matching icon out.

n8n nodes for CalDAV calendars: events, tasks, free/busy, and a polling trigger for Apple iCloud, Google, Nextcloud, Fastmail, Radicale, Baikal, and any RFC 4791 server

Bounded Google Tag Manager API v2 nodes and role-separated credentials for n8n

Embeddable React AI chatbot widget with streaming responses, authentication, Markdown rendering, and conversation support.