color-switch
ChatGPTUse to launch the Color Switch brain-training game as an interactive widget inside the ChatGPT conversation. Call this tool when the user wants to play, practice, test themselves, warm up mentally, or try a short challenge related to switching attention between written words and ink colors while resisting the wrong automatic answer. Typical requests include prompts such as “give me a Stroop test,” “open a color-word focus game,” or “let’s do an attention control challenge”. This tool is for active participation: the user is expected to interact with the widget directly, complete a timed session, and then continue the conversation about the result. When this tool is invoked, ChatGPT renders the Color Switch widget rather than returning a text-only explanation. The widget handles the countdown, the one-minute round, live progress display, scoring, final result presentation, and persistence of the completed result in widget state. Inside the widget, the user sees a color word in a potentially conflicting ink color, follows the visible current rule, and taps one of four answer buttons to respond. The score increases for correct prompts and decreases for wrong or expired prompts, using (successfulPrompts - failedPrompts) * 100. After the game ends, the widget can send a follow-up prompt asking the assistant to analyze the result and offer a sensible next step, such as replaying the same game or trying Pattern Recall. Prefer this tool when the user clearly wants a playable mini-game, an interactive cognitive exercise, or a score they can discuss with the model afterward. It is also appropriate when the user asks for a number game, a visual spotting game, a reaction-speed style challenge, or a mental dexterity task that is better experienced interactively than described abstractly. If the user is comparing different game types, choose this tool only when switching attention between written words and ink colors while resisting the wrong automatic answer is the best match for what they asked for. Do not call this tool when the user explicitly asks for plain text only, asks not to use tools, wants general advice about focus or cognition without playing a game, or is already discussing a completed score and only needs analysis. If the user instead wants a tap-back memory sequence, the pattern-recall tool is the better fit. If they want a spatial tracing challenge, the path-trace tool is the better fit. Also avoid calling it for unrelated casual conversation, factual questions, or tasks that do not benefit from an embedded widget. In short, this tool is for situations where an interactive, measurable, one-minute brain-training task will help the user engage directly and then reflect with the assistant. It should be the default choice only when the user wants to play Color Switch, or when their request strongly implies a game centered on switching attention between written words and ink colors while resisting the wrong automatic answer, with the expectation of an immediate score and a follow-up conversation about performance.
number-sort
ChatGPTUse to launch the Number Sort brain-training game as an interactive widget inside the ChatGPT conversation. Call this tool when the user wants to play, practice, test themselves, warm up mentally, or try a short challenge related to ordering numbers accurately under time pressure. Typical requests include prompts such as “let’s play a number game,” “give me a brain-training sorting challenge,” or “test my speed with numbers”. This tool is for active participation: the user is expected to interact with the widget directly, complete a timed session, and then continue the conversation about the result. When this tool is invoked, ChatGPT renders the Number Sort widget rather than returning a text-only explanation. The widget handles the countdown, the one-minute round, live progress display, scoring, final result presentation, and persistence of the completed result in widget state. Inside the widget, the user sees three numbered targets on a large grid and must tap them from lowest to highest without making a mistake. The final score is derived from successful rounds minus failed rounds, multiplied by one hundred, which rewards both pace and correctness. After the game ends, the widget can send a follow-up prompt asking the assistant to analyze the result and offer a sensible next step, such as replaying the same game or trying Odd One Out. Prefer this tool when the user clearly wants a playable mini-game, an interactive cognitive exercise, or a score they can discuss with the model afterward. It is also appropriate when the user asks for a number game, a visual spotting game, a reaction-speed style challenge, or a mental dexterity task that is better experienced interactively than described abstractly. If the user is comparing different game types, choose this tool only when ordering numbers accurately under time pressure is the best match for what they asked for. Do not call this tool when the user explicitly asks for plain text only, asks not to use tools, wants general advice about focus or cognition without playing a game, or is already discussing a completed score and only needs analysis. If the user instead asks for a visual oddity game, shape spotting, or orientation matching, the odd-one-out tool is the better fit. Also avoid calling it for unrelated casual conversation, factual questions, or tasks that do not benefit from an embedded widget. In short, this tool is for situations where an interactive, measurable, one-minute brain-training task will help the user engage directly and then reflect with the assistant. It should be the default choice only when the user wants to play Number Sort, or when their request strongly implies a game centered on ordering numbers accurately under time pressure, with the expectation of an immediate score and a follow-up conversation about performance.
odd-one-out
ChatGPTUse to launch the Odd One Out brain-training game as an interactive widget inside the ChatGPT conversation. Call this tool when the user wants to play, practice, test themselves, warm up mentally, or try a short challenge related to finding the single rotated item in a field of matching emojis. Typical requests include prompts such as “let’s play a visual brain game,” “give me an odd one out challenge,” or “test my visual spotting speed”. This tool is for active participation: the user is expected to interact with the widget directly, complete a timed session, and then continue the conversation about the result. When this tool is invoked, ChatGPT renders the Odd One Out widget rather than returning a text-only explanation. The widget handles the countdown, the one-minute round, live progress display, scoring, final result presentation, and persistence of the completed result in widget state. Inside the widget, the user scans a 5 by 5 emoji grid and taps the one cell whose orientation differs from the rest. Every successful round increases the final score by one hundred points, while wrong taps count as failed rounds and waste time. After the game ends, the widget can send a follow-up prompt asking the assistant to analyze the result and offer a sensible next step, such as replaying the same game or trying Number Sort. Prefer this tool when the user clearly wants a playable mini-game, an interactive cognitive exercise, or a score they can discuss with the model afterward. It is also appropriate when the user asks for a number game, a visual spotting game, a reaction-speed style challenge, or a mental dexterity task that is better experienced interactively than described abstractly. If the user is comparing different game types, choose this tool only when finding the single rotated item in a field of matching emojis is the best match for what they asked for. Do not call this tool when the user explicitly asks for plain text only, asks not to use tools, wants general advice about focus or cognition without playing a game, or is already discussing a completed score and only needs analysis. If the user instead asks for ordering numbers, tapping a sequence, or a numeric speed challenge, the number-sort tool is the better fit. Also avoid calling it for unrelated casual conversation, factual questions, or tasks that do not benefit from an embedded widget. In short, this tool is for situations where an interactive, measurable, one-minute brain-training task will help the user engage directly and then reflect with the assistant. It should be the default choice only when the user wants to play Odd One Out, or when their request strongly implies a game centered on finding the single rotated item in a field of matching emojis, with the expectation of an immediate score and a follow-up conversation about performance.
path-trace
ChatGPTUse to launch the Path Trace brain-training game as an interactive widget inside the ChatGPT conversation. Call this tool when the user wants to play, practice, test themselves, warm up mentally, or try a short challenge related to remembering a short route across a grid and reproducing it step by step. Typical requests include prompts such as “open a path memory game,” “test my spatial recall,” or “give me a route tracing challenge”. This tool is for active participation: the user is expected to interact with the widget directly, complete a timed session, and then continue the conversation about the result. When this tool is invoked, ChatGPT renders the Path Trace widget rather than returning a text-only explanation. The widget handles the countdown, the one-minute round, live progress display, scoring, final result presentation, and persistence of the completed result in widget state. Inside the widget, the user studies a short numbered route on a grid, waits for it to disappear, then retraces the same adjacent path with taps or a drag gesture. Every successful retrace adds one round of progress, every wrong path subtracts one, and the score is shown as (successfulRounds - failedRounds) * 100. After the game ends, the widget can send a follow-up prompt asking the assistant to analyze the result and offer a sensible next step, such as replaying the same game or trying Word Run. Prefer this tool when the user clearly wants a playable mini-game, an interactive cognitive exercise, or a score they can discuss with the model afterward. It is also appropriate when the user asks for a number game, a visual spotting game, a reaction-speed style challenge, or a mental dexterity task that is better experienced interactively than described abstractly. If the user is comparing different game types, choose this tool only when remembering a short route across a grid and reproducing it step by step is the best match for what they asked for. Do not call this tool when the user explicitly asks for plain text only, asks not to use tools, wants general advice about focus or cognition without playing a game, or is already discussing a completed score and only needs analysis. If the user instead wants a letter-grid spelling task, the word-run tool is the better fit. If they want a tap-back pattern memory sequence without dragging, the pattern-recall tool is the better fit. Also avoid calling it for unrelated casual conversation, factual questions, or tasks that do not benefit from an embedded widget. In short, this tool is for situations where an interactive, measurable, one-minute brain-training task will help the user engage directly and then reflect with the assistant. It should be the default choice only when the user wants to play Path Trace, or when their request strongly implies a game centered on remembering a short route across a grid and reproducing it step by step, with the expectation of an immediate score and a follow-up conversation about performance.
pattern-recall
ChatGPTUse to launch the Pattern Recall brain-training game as an interactive widget inside the ChatGPT conversation. Call this tool when the user wants to play, practice, test themselves, warm up mentally, or try a short challenge related to remembering and replaying a short sequence of highlighted grid positions. Typical requests include prompts such as “give me a memory sequence game,” “open a Simon-style challenge,” or “test my pattern recall”. This tool is for active participation: the user is expected to interact with the widget directly, complete a timed session, and then continue the conversation about the result. When this tool is invoked, ChatGPT renders the Pattern Recall widget rather than returning a text-only explanation. The widget handles the countdown, the one-minute round, live progress display, scoring, final result presentation, and persistence of the completed result in widget state. Inside the widget, the user watches a short grid sequence light up and then taps the same cells back in order from memory. The score is derived from successful minus failed rounds, multiplied by one hundred, so a clean recall streak matters more than random speed. After the game ends, the widget can send a follow-up prompt asking the assistant to analyze the result and offer a sensible next step, such as replaying the same game or trying Path Trace. Prefer this tool when the user clearly wants a playable mini-game, an interactive cognitive exercise, or a score they can discuss with the model afterward. It is also appropriate when the user asks for a number game, a visual spotting game, a reaction-speed style challenge, or a mental dexterity task that is better experienced interactively than described abstractly. If the user is comparing different game types, choose this tool only when remembering and replaying a short sequence of highlighted grid positions is the best match for what they asked for. Do not call this tool when the user explicitly asks for plain text only, asks not to use tools, wants general advice about focus or cognition without playing a game, or is already discussing a completed score and only needs analysis. If the user instead wants a drag-based route memory challenge, the path-trace tool is the better fit. If they want an inhibition task built around colors and words, the color-switch tool is the better fit. Also avoid calling it for unrelated casual conversation, factual questions, or tasks that do not benefit from an embedded widget. In short, this tool is for situations where an interactive, measurable, one-minute brain-training task will help the user engage directly and then reflect with the assistant. It should be the default choice only when the user wants to play Pattern Recall, or when their request strongly implies a game centered on remembering and replaying a short sequence of highlighted grid positions, with the expectation of an immediate score and a follow-up conversation about performance.
word-run
ChatGPTUse to launch the Word Run brain-training game as an interactive widget inside the ChatGPT conversation. Call this tool when the user wants to play, practice, test themselves, warm up mentally, or try a short challenge related to forming valid words from adjacent letters under time pressure. Typical requests include prompts such as “let’s play a word game,” “give me a spelling challenge,” or “open a letter-grid brain training game”. This tool is for active participation: the user is expected to interact with the widget directly, complete a timed session, and then continue the conversation about the result. When this tool is invoked, ChatGPT renders the Word Run widget rather than returning a text-only explanation. The widget handles the countdown, the one-minute round, live progress display, scoring, final result presentation, and persistence of the completed result in widget state. Inside the widget, the user works across one fixed 5 by 5 letter board, taps or drags through neighboring letters, including diagonals, to spell a word, and taps the latest selected letter again to submit that spelling attempt. Every rejected submission costs one hundred points, while every accepted word awards one hundred points times the word length minus two, rewarding longer correct words rather than simple guess volume. After the game ends, the widget can send a follow-up prompt asking the assistant to analyze the result and offer a sensible next step, such as replaying the same game or trying Number Sort. Prefer this tool when the user clearly wants a playable mini-game, an interactive cognitive exercise, or a score they can discuss with the model afterward. It is also appropriate when the user asks for a number game, a visual spotting game, a reaction-speed style challenge, or a mental dexterity task that is better experienced interactively than described abstractly. If the user is comparing different game types, choose this tool only when forming valid words from adjacent letters under time pressure is the best match for what they asked for. Do not call this tool when the user explicitly asks for plain text only, asks not to use tools, wants general advice about focus or cognition without playing a game, or is already discussing a completed score and only needs analysis. If the user instead asks for number ordering, sequence tapping, or a pure numeric speed challenge, the number-sort tool is the better fit. If they want a visual oddity or rotated-emoji spotting task, the odd-one-out tool is the better fit. Also avoid calling it for unrelated casual conversation, factual questions, or tasks that do not benefit from an embedded widget. In short, this tool is for situations where an interactive, measurable, one-minute brain-training task will help the user engage directly and then reflect with the assistant. It should be the default choice only when the user wants to play Word Run, or when their request strongly implies a game centered on forming valid words from adjacent letters under time pressure, with the expectation of an immediate score and a follow-up conversation about performance.
