angshuman-nandy/LLM-Battleships

added gpt 5.5 and prompt updates

Revert "use websockets instead of sse and prompt updates"

use websockets instead of sse and prompt updates

updated hit logic in prompt

added pushover notification

removed ollama

Default board size to 5x5; restrict to Haiku/GPT-4o-mini only

keeping only the cheap models2

keeping only the cheap models

fixes to prompts

add .env example

Add MIT license headers to all source files and root LICENSE file

Prevent LLM from repeating shots: add fired-cell list to prompt and validate in Anthropic wrapper

Restrict model choices to curated presets, remove custom model input

Fix LLM placement: explicit bounds hints and random fallback after retries

Replace deprecated claude-3-5-sonnet-20241022 with claude-opus-4-7 preset

Add model preset dropdown with custom model option to SetupPanel

updates

Add Hugging Face Spaces metadata to README

init commit