Opentable logo

Opentable

Community
win4r
opentable

Find restaurants on OpenTable, check availability, and make, change, or cancel reservations. Use for restaurant booking and live reservation data.

Overview

Publisherwin4r
RepositoryMuseAI-Skills
Skill nameopentable
Stars
382
Forks
116
Bundled files
2
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by win4r on GitHub. Read the source before you install it.

Installation

Install the Opentable AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/win4r/MuseAI-Skills/tree/main/opt/hatch/skills/opentable
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/win4r/MuseAI-Skills.git /tmp/MuseAI-Skills
mkdir -p .claude/skills
cp -r /tmp/MuseAI-Skills/opt/hatch/skills/opentable .claude/skills/opentable
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Opentable in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Opentable on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Opentable is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

OpenTable

For a user-facing restaurant search or booking, first read /opt/hatch/skills/booking/SKILL.md, /opt/hatch/skills/booking/references/restaurants.md, and /opt/hatch/skills/booking/references/presentation.md. Those files define the end-to-end booking, fallback, and presentation rules. Use this file for the OpenTable CLI contract. Keep reservation choices in plain Markdown. Do not call create_options or a comparison/list widget during a restaurant booking.

Connecting

OpenTable needs a one-time in-chat consent before any command returns data. Run opentable status. For a direct reservation request, if it is not_connected, recommend connecting OpenTable as the preferred low-friction path and post the exact returned connect_url. Do not invent one. Explain briefly that this enables live availability, saved-profile use, and smoother booking. Do not wait idle for the connection or make it a prerequisite: continue the same request through the venue's official reservation link, another reputable platform, and phone.place_call when available. If the user connects after a fallback path started, follow the duplicate-protection rule in /opt/hatch/skills/booking/references/restaurants.md before resuming the OpenTable request. If status is unavailable, skip the connection offer and use those fallback paths immediately. If the request is specifically to connect OpenTable rather than reserve a table, post the link immediately and wait for the user to connect. Disconnect with opentable disconnect. Do not send the user to Settings.

Common flows

Find a table

Use lookup-rid to resolve a named restaurant or discover restaurants by city, cuisine, and price. When the city is known, include --city. If there is no clear match, vary the restaurant name or adjust the filters. For example, try common spacing or punctuation variants, pass only the city name to --city, or drop --country-code. Keep the requested location fixed.

Then use search-availability to check open tables for the requested time and party size.

Book

After search-availability, resolve the exact slot from the user's request. Use authorized stored contact details when available; ask only for the name, email, or phone fields still required at checkout. Then book with book-reservation (it locks the slot and books in one step). After book-reservation returns a confirmation, offer to remind the user 30 minutes before the reservation. If they accept or choose another lead time, use cron.add to create a run-once reminder for their chosen time, resolved from the confirmed reservation time.

Modify

Find the new time with search-availability, resolve the exact requested slot, then update with modify-reservation-with-lock. Needs the reservation's confirmation id.

Cancel or look up a reservation

Use cancel-reservation or get-reservation, both by the reservation's confirmation id.

Book an experience

Find it with list-experiences, check times with search-availability --include-experiences true, resolve the exact requested experience and slot, then book with book-reservation --experience-json (include the experience id and version).

Other commands

The flows above cover the common cases. For anything else (a restaurant's policies, seating and dining-area options, releasing a stuck slot lock), run opentable --help for the full command list and opentable <command> --help for its flags.

Rules

  • Before book-reservation or modify-reservation-with-lock, resolve the exact restaurant, date, time, and party size. Before cancel-reservation, resolve which reservation the user means. For a booking, take the user's name, email, and phone from an authorized profile or the user. Invoke the resolved command directly so connector policy can present any required approval. Do not add a duplicate chat confirmation. A clear, unambiguous cancellation may proceed without an additional confirmation. Do not guess missing details.
  • Read results add UTC and user-local semantic fields when OpenTable supplies an offset-bearing reservation or hold timestamp, plus runtime-generated retrieved_at. A date/time without an offset is restaurant-local civil time; do not guess a timezone or convert it.
  • book-reservation verifies the booking with OpenTable before reporting success. Only report the booking as confirmed when the result explicitly says it is confirmed; a confirmation number alone is not proof. When it is confirmed, say so without hedging, mentioning provider lifecycle states, or predicting a later status transition. If OpenTable requires another action, explain what is needed instead of retrying. If the booking is unconfirmed or cannot be verified, say that plainly, keep the confirmation reference, and never retry automatically. Offer only the OpenTable continuation or recovery link returned by the command; the link itself never proves confirmation, and you must never invent one.
  • For reservation lookups, summarize the returned status in a natural sentence (for example, “Your reservation is confirmed”). Never quote result field names or provider-internal status fields. Describe cancelled, completed, no-show, or unknown results in the same plain language.
  • When OpenTable returns a reservation-management link, include it exactly once as the only web link in the final reply, on its own last line. Do not also link the restaurant profile; multiple links prevent Hatch from rendering the management card.
  • Only tell the user a booking, change, or cancellation went through when the command returns a confirmation. If it fails or comes back empty, say so plainly instead of inventing a confirmation or a workaround.
  • A result carrying resource_authorization with persisted: false means the OpenTable operation succeeded but the main agent could not confirm that its authorization was stored. Do not automatically repeat a mutation. Give the user the confirmation number and explain that the main agent may not be able to manage it later. Reads are safe to retry except after a rate-limit response.
  • Make only one OpenTable connector call at a time. Do not batch calls, run them in parallel or in the background, or put them in shell/Python loops or retry wrappers.
  • If an OpenTable command returns HTTP 429 or says it was rate limited, stop making OpenTable calls for this task and report the partial result. Never repeat book-reservation, modify-reservation-with-lock, or cancel-reservation after a rate limit because the mutation may already have applied. If the result includes a confirmation reference, a later task may retry only get-reservation after the returned retry_after, or after 60 seconds if the response has no value.
  • A reservation the main agent did not book may not be reachable by get-reservation, modify-reservation-with-lock, or cancel-reservation. If one of those commands is refused, say so and point the user to opentable.com or the restaurant. Do not retry with a different rid or confirmation id.
  • Attribute each reservation task to OpenTable. Mention OpenTable when presenting availability. Mention it once more in the final result only when that clarifies who owns the reservation. Do not repeat it in intermediate updates or use promotional language.
  • Only add special requests the user gave you (dietary needs, allergies, seating). Do not put unrelated personal data in the booking.
  • If lookup-rid finds no match after the applicable retries, stop the OpenTable attempt without checking availability or inventing a rid. If the user asked only whether the venue appears on OpenTable, report that result. If the user intends to reserve, continue with the Search and route section in /opt/hatch/skills/booking/references/restaurants.md.
  • If search-availability returns no_availability_reasons, tell the user why in plain language rather than showing the raw code.
  • Present restaurant results through /opt/hatch/skills/booking/references/presentation.md.
  • Keep messages focused on the outcome. Do not quote the command you ran or plan to run, rids, tokens, slot-lock state, or raw result codes like not_connected or NoTimesExist. Tell the user the outcome in normal words. The OpenTable attribution described above is user-facing. It is not an internal implementation detail.

Limits

  • The connector cannot book a slot that needs a card, deposit, or prepayment. These show up as a cancellation_policy on the slot in search-availability, or an experience marked prePaymentRequired. When these requirements are present, or a booking is rejected for needing a card, continue the same reservation with browser.spawn_task following /opt/hatch/skills/booking/references/browser-booking.md. Use the returned booking_url with ref=19075 added if missing, or the restaurant's profile_url if no booking URL is returned.
  • Bookings are for 1 to 20 diners. For a larger party, tell the user to arrange it with the restaurant directly. Do not book a smaller table or split the group to fit.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Opentable AI skill do?

Find restaurants on OpenTable, check availability, and make, change, or cancel reservations. Use for restaurant booking and live reservation data.

Why use Opentable on TypingMind?

Because you install it once and use it with any model. Opentable is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Opentable in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/win4r/MuseAI-Skills/tree/main/opt/hatch/skills/opentable. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Opentable?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Opentable?

As many as you like. As long as a model supports skills, you can use Opentable with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Opentable AI skill free?

It is published on GitHub by win4r. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇