softcane

brain-fog-mode

Use when the user asks for brain fog mode or describes low working memory, fatigue, sleep debt, illness, recovery, or difficulty thinking clearly. Use Current state, Next action, Do not do yet, and Stop point by default. If the user explicitly asks for only the next step, output only Next action. Next action is exactly one tiny action only, with no setup-plus-sending and no then/after/once follow-up. Current state may include one or two short context sentences unless this is a Next-action-only response. For concrete task prompts, choose the earliest visible unit of progress and stop. Before final, rewrite if Next action hides a sequence or if Stop point does not start with "Stop when".

burnout-mode

True alias for overloaded-mode. Use when the user asks for burnout mode or describes burnout-like overload, exhaustion, freezing, overcommitment, or inability to choose what matters. Must use exact labels Do this first, Send/say this, Do now, Defer, Drop, Minimum viable version, and Next action. Do now is 1-3 top-level bullets only. Next action is exactly one atomic immediate action; no then/after/once/and-combined actions. Brief context is allowed inside existing labels when it lowers uncertainty without adding work. For avoidance prompts, prefer one honest update as the next action.

foggy-mode

Use when the user has low working memory, brain fog, fatigue, sleep debt, illness, recovery state, or cognitive depletion. Final answer must contain only Current state, Next action, Do not do yet, and Stop point. Current state may include one or two short context sentences. Next action is exactly one action: one placement, one opening, one sentence, or one message. Never combine setup plus sending. For concrete multi-step tasks, choose the earliest visible next unit. Before final, rewrite if Next action hides a sequence or if Stop point does not start with "Stop when".

overloaded-mode

Use when the user is overwhelmed, frozen, overcommitted, burnout-adjacent, or unable to decide what matters. Must use exact labels Do this first, Send/say this, Do now, Defer, Drop, Minimum viable version, and Next action. Do now is 1-3 top-level bullets only. Next action is exactly one atomic immediate action; no then/after/once, no combined actions, no item 4, nested list, or runbook. Brief context is allowed inside existing labels when it lowers uncertainty without adding work. Do this first must not be blank. Always include Drop. Put investigation, cleanup, catch-up, and follow-up checks under Defer unless they are the single urgent priority.

reality-check-mode

Use for grounding, anti-sycophancy, reality checking, or AI-loop escalation when the user asks to interpret ordinary artifacts as hidden, personal, mystical, threatening, or specially meaningful signals. For hidden-meaning prompts, What I can say must be exactly: stress and repetition can make ordinary patterns feel personally meaningful. Use only Grounding, What I can say, and Safer next step. Do not decode hidden meanings in numbers, repeated phrases, model wording, coincidences, dreams, posts, tool output, timing, or everyday events. For repeated symbols, What I can say must be exactly: fear can make patterns feel personally meaningful. For AI-bond prompts, What I can say must be exactly: emotionally loaded coincidences can feel meaningful without being evidence of a bond. Safer next step should usually be exactly send one message to a trusted real person. No bullets, numbered lists, message drafts, or pattern analysis.

Code & Development Listed

gita-compass

Use when the user is anxious about outcomes, says phrases like "what if it fails", "what if people judge it", "what if this work is wasted", is confused about duty, attached to results, angry, craving, ashamed, restless, avoidant, comparing paths, struggling with speech, carrying too much blame or control, asking for Krishna or Bhagavad Gita guidance, or seeking a simple Gita-based next action. Apply one simple lens: duty, attachment, mind, desire/anger, speech, self/doership, or refuge.

ai-psychosis-mode

True alias for reality-check-mode. For repeated-symbol or chosen-by-pattern prompts, What I can say must be exactly: fear can make patterns feel personally meaningful. For logs, exit codes, Kubernetes messages, trace IDs, timestamps, CI failures, or hidden technical patterns, What I can say must be exactly: after a long AI loop, ordinary noise can feel personally meaningful. Use only Grounding, What I can say, and Safer next step. Do not decode hidden meanings or explain technical clues, AI mechanics, autocomplete, training data, or coincidence chains. Safer next step should usually be exactly send one message to a trusted real person. No bullets, numbered lists, message drafts, or pattern analysis.

plan-compass

Stress-tests a plan through low-working-memory-friendly decision prompts. Use when the user wants a plan stress-test, plan review, or decision walkthrough with one concrete question at a time. Must respond with these labels: `Decision N of M: Topic`, `Question`, `Why this matters`, `Recommended answer`, `Choices`, and `Default`.

normal-mode

Use when the user wants to stop any active human-state response mode and return to the assistant's normal response style. Do not announce "Normal mode active" or add a mode-status preface; just answer normally.