AI Agent Use Cases
Views —

AI Agent Use Cases

Added today

44All 189

Prompt to Prepare a Shopping Cart from a Recipe Photo

The author reports that Grok Bot receives a recipe photo, organizes the ingredients and quantities, and creates a shopping list by excluding ingredients already available at home.

Grok Bot
Life & Hobbies

1–20 of 189

Life & HobbiesWalkthrough

Agent That Reads Chessboard State and Makes the Next Move

This is a public demo where a chess agent reads the board state, selects the next move, submits it, and then waits for its turn again.

Claude Code

How it works

  1. The author prepares the ChessAI game server and a Claude Code sub-agent to start a test match.
  2. The Claude sub-agent repeatedly reads the board state via HTTP, selects a move, posts it, and then waits for its next turn.

Results

In the public example, the intended outcome is a chess opponent that makes moves based on the board state.

Notes

This is a public game demo. The game code separately determines whether moves are legal, while the model selects the next move.

Source

ChessAI 공개 저장소 Original post ↗

Example prompt · Claude Code

Goal: Create an opponent that reads the changed chessboard and makes the next move.

Input/Environment: [Public repository], [Test server], [Claude Code environment], [Initial chessboard].

Reference original: https://github.com/devasur/chessai

Prepare the configuration and connections so that a test environment can be created using the public setup from the original. If there are any missing sub-workflows, APIs, or account permissions, mark them as incomplete until they are actually connected.

Connections and materials to prepare: Prepare the ChessAI game server and a Claude Code sub-agent to start a test match.

Agent behavior to implement: The Claude sub-agent should repeatedly read the board state via HTTP, select a move, post it, and then wait for its next turn.

Scope of operation: This is a public game demo. The game code separately determines whether moves are legal, while the model selects the next move.

The result should be a chess opponent that makes moves based on the board state. Please also organize the necessary setup steps and the actual execution method. After checking the results returned by the tool, choose the next task, and separately note any resources that cannot be accessed or tasks that could not be executed.

Verification: Check with actual game records whether the agent refrains from making moves when it is not its turn, handles rejected moves, and processes game-over states. Show the input used for verification, the actual output, and clearly distinguish any items that have not yet been confirmed.