← iPlugin CheckersCONTENT HISTORY

Update to iPlugin Checkers

Snapshot Nov 18, 2026 · 06:17 UTC · version 0.1.6

Collection source: downloaded plugin package.

WHAT CHANGED

First saved snapshot

No earlier snapshot is available to establish a change.

Save iPlugin Checkers and check future changes in Following.

Up to 3 plugins free · no card required.

Compare saved observations

Download comparison JSON
Full technical diff · 0 changed fields
Full snapshot data
{
  "description": "Use when the user wants to play Chinese checkers (Sternhalma), open the fullscreen star board, make a move, jump, inspect a game, undo, or resign.",
  "included_files": [],
  "name": "checkers",
  "skill_md_contents": "---\nname: checkers\ndescription: Use when the user wants to play Chinese checkers (Sternhalma), open the fullscreen star board, make a move, jump, inspect a game, undo, or resign.\n---\n\n# iPlugin Checkers\n\nUse this skill when the user asks to play Chinese checkers, 跳棋, marbles, the star board, or wants a 1v1 board game of stepping and jumping. Call `checkers_home` **once** to open the fullscreen board. It is the only tool that renders UI; keep every game action and model reply inside that same board. Never call `checkers_home` again after the board is open, not even to start a new game: call `checkers_start_game` in the existing view. Extra home calls create extra app tabs.\n\nThe game is a 1v1 Chinese checkers race on a 121-hole star board. The human always plays the ten **porcelain** pegs from the bottom camp. The model always plays the ten **cinnabar** pegs from the top camp. `firstPlayer` only decides who takes the first turn; it does not change camps.\n\n## Goal and turn rules\n\n- The goal is to move all ten of your pegs into the ten holes of the **opposite camp**. The porcelain side wins by filling the top camp; cinnabar wins by filling the bottom camp.\n- A turn is exactly one of:\n  - one **step** to an adjacent empty hole, or\n  - one or more **jumps**: leap over one adjacent peg (either colour) into the empty hole immediately beyond, then optionally continue jumping from the landing hole in any direction.\n- A jump chain is applied as one move. The server chooses and validates the complete path from the `from` and `to` holes.\n- Pegs are never captured. A jumping peg merely passes over another peg.\n- A peg may not land on an occupied hole, and may not combine a step with a jump in the same turn.\n\n## Coordinates\n\nRows are lettered `A` through `Q` from the top point to the bottom point. Within each row, holes are numbered left to right. Because the star rows have different widths, valid numbers depend on the row (`E1`–`E13` on the wide row E, for example). Always prefer coordinates from the latest `legalMoves` list instead of guessing a coordinate.\n\n## Tools\n\n- `checkers_home` — open the fullscreen board. Accepts no arguments.\n- `checkers_start_game` — start a 1v1 game with `firstPlayer` (`human` or `ai`) and `difficulty` (`easy`, `normal`, or `hard`).\n- `checkers_get_state` — read the authoritative board, pieces, legal moves, progress, history, and next turn for a `gameId`.\n- `checkers_play_move` — apply one legal human move using `gameId`, `from`, and `to`.\n- `checkers_ai_move` — apply the model's chosen cinnabar move using `gameId`, `from`, and `to`. The server rejects any coordinate that is not in the current legal move list.\n- `checkers_undo` — permanently remove the latest complete turn (the cinnabar reply plus the human move, a pending human move, or a cinnabar opening). Removed moves cannot be restored.\n- `checkers_resign` — end the active game after the human explicitly confirms resignation.\n\n## Playing as the model\n\nThe model is the cinnabar player. There is no engine, opening book, learning system, or hidden move selector. After every human move, or after starting with `firstPlayer: \"ai\"`, inspect the returned `legalMoves` and immediately call `checkers_ai_move` with one of them. Do not reply with text only and do not leave the cinnabar turn waiting.\n\nPrefer moves that:\n- advance pegs out of the crowded home camp,\n- use jumps and multi-jump chains to cover distance,\n- open lanes for the pegs still at home,\n- block the human's shortest route when the position calls for it.\n\nHonour the selected difficulty:\n- `easy`: favour short, simple moves and avoid long forcing jump chains.\n- `normal`: balance jump distance, lane development, and home-camp clearance.\n- `hard`: look for the longest useful jump chains, efficient camp evacuation, and blocking moves.\n\nThe state reports `progress.human` and `progress.ai` as the number of pegs already in the far camp (0–10). The game ends at 10. If the state is `finished` or `draw`, do not attempt another move; explain the result briefly and offer a new game. Before resignation, ask the user to confirm.\n"
}

SHA-256 of public snapshot: ce4116c40b4f5f39f8c61a7459dc4cdd0d6bb417eb266a1976dc7006b6c5ec71