An intelligent, image-enabled AI sommelier that evaluates beer and wine selections against a dynamically updated profile of your personal taste preferences.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install drinks-sommelier
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install drinks-sommelier using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Drinks Sommelier?
The drinks-sommelier is a specialized module designed to transform your AI assistant into an expert virtual sommelier. Operating within the Openclaw Skills framework, this skill enables agents to parse text lists, analyze shelf or menu images, and perform targeted real-time web searches to evaluate beers and wines. Rather than providing generic recommendations, it builds an increasingly precise, multi-tiered profile of your taste rules and historical preferences over time.
By cross-referencing qualitative rules (such as bitterness limits or preferred sweetness levels) with records of specific products you like or dislike, the skill delivers highly customized suggestions. It ensures that every recommendation is backed by up-to-date online data rather than outdated training sets, making it a reliable companion for navigating physical shelves, digital cellars, or complex restaurant menus.
Drinks Sommelier Use Cases
- Analyzing a photo of a crowded wine shelf at a local retailer to isolate the highest-rated bottle matching your flavor profile.
- Parsing a craft beer menu at a brewery to highlight options matching your specific bitterness (IBU) and alcohol content limits.
- Selecting the ideal beverage pairing for a meal based on text lists of available inventory.
- Gradually refining a persistent local dataset of liked and disliked drinks to continuously train your personal AI agent.
How Drinks Sommelier Works
- Initialization Check: The skill verifies whether initial taste rules for beers and wines have been defined in its configuration. If not, it guides you through a setup process.
- Profile Loading: It reads persistent local Markdown files (
known-preferred-beers.mdandknown-preferred-wines.md) along with user taste rules to build a multi-level mental model of your preferences. - Input Processing: It processes user inputs containing text names, images (menus, shelves, lists), or a mixture of both, filtering out irrelevant items like food or spirits.
- Targeted Information Retrieval: For identified items, it executes live web queries (using browser-search skills or native tools) to fetch current specifications such as grape variety, IBU, ABV, and regional characteristics.
- Preference Scoring: It evaluates each option, assigning a preference index from 0% to 100% based on how closely the specifications align with your taste rules.
- Response Construction: It outputs a structured recommendation featuring first choices, indexed alternatives, food pairings, and items to explicitly avoid.
- Continuous Synchronization: Feedback on suggestions (e.g., "I didn't like this wine") triggers automatic, local file updates to refine future runs.
Drinks Sommelier Setup
To deploy the drinks-sommelier skill, clone or download the repository into your custom directory and perform the initial setup:
# Navigate to your skill directory
cd path/to/skills/drinks-sommelier
# Review and update setup files
cat data/SETUP.md
Next, initialize the main configuration in SKILL.md by replacing the default placeholder text Enter your preferences here under the User's tastes for beers and User's tastes for wines sections with your qualitative rules. Ensure the following files exist in your local workspace to track specific product feedback:
data/known-preferred-beers.mddata/known-preferred-wines.md
Drinks Sommelier Data Schema & Taxonomy
The drinks-sommelier skill organizes persistent profile data within local Markdown files to build a private knowledge graph of your taste preferences:
| File Path | Content Type | Operational Role |
|---|---|---|
SKILL.md |
Taste Rules (Qualitative) | High-priority binding rules (e.g., maximum ABV, bitterness tolerance) that every recommendation must satisfy. |
data/known-preferred-beers.md |
Liked & Disliked Beers | Specific historical beer logs used to calibrate the qualitative rules. |
data/known-preferred-wines.md |
Liked & Disliked Wines | Specific historical wine logs used to calibrate the qualitative rules. |
data/SETUP.md |
Initial Onboarding Guidelines | Instructions for guiding the initial configuration of user taste profiles. |
Drinks Sommelier Advanced Features
- Image-Based Object Recognition: Leverages visual models to extract product names directly from complex wine shelves, menus, and refrigerator displays.
- Web Search Integration: Uses active browser search tools to fetch real-time specifications, avoiding reliance on outdated pre-training knowledge.
- Dynamic Feedback Loop: Automatically updates the local data files as you interact with the agent, bypassing manual file edits for standard feedback.
- Two-Level Preference Calibration: Distinguishes between strict taste rules and concrete product reviews, resolving conflicts between general statements and specific products.
- Proactive Conflict Resolution: Flags discrepancies (e.g., if you dislike a wine that matches your expressed rules) and prompts for clarification to maintain data integrity within Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Makaron CLI Creative Agent
A CLI tool and agent skill to dynamically generate, edit, and assemble images, video, music, and motion designs.
iOS CocoaPods to SPM Migration
A structured workflow and prompt suite for migrating iOS projects from CocoaPods to Swift Package Manager.
chatmosp-file-organizer
A secure file system manager for the chatMOSP system that automates task directory creation, intelligent naming formats, and secure file operations.
chatmosp-input-coordinator
An entry skill for ChatMOSP that processes multilingual inputs to extract parameters and coordinate workflows for metal structure generation and kinetic simulations.
Browser Search
A multi-layered web search and browsing engine that scales from lightweight SearXNG queries to advanced CloakBrowser anti-bot evasion.
Break Watch
A Python-powered A-share stock market scanner designed to identify real-time breakout patterns combining volume surges, upward price movements, and bullish candles.
Featured*
