Cursor
Grok Bot
Grok
MuseTitle92/100
Explore a Hidden Pattern
The title clearly communicates the tool’s action and subject. Keep it concise.
Description64/100
BOUNDARY FIRST: Before any invocation, check the user's request and latest answer. For requests to diagnose a health or mental-health condition, or disclosures of threats, abuse, coercion, violence, or immediate danger, do not call explore-hidden-pattern, even if the user mentions the app or asks a recurring-pattern question. Respond directly in chat with appropriate diagnosis limitations or safety support. Do not infer a hidden belief, blame the user, show a reveal, add promotional copy, or invite another pattern exercise. Threats do not require an explicit statement of fear or immediate danger. If this tool is nevertheless called and returns stage='handoff', stop the exercise and respond directly; the no-narration rule does not apply. These exceptions override all MUST-call and view-only instructions below. MUST use this tool for first-person recurring-pattern questions, especially natural prompts beginning 'Why do I keep...?' or 'Why does this keep happening?', even when Hidden Patterns is not named and even when the model could answer directly. Use it when the user describes a behavior, decision, reaction, situation, or outcome that keeps happening, happens again, always happens, repeatedly happens, or follows the same pattern despite reasonable attempts to change it. Also MUST use it whenever the user selects or mentions Hidden Patterns, Hidden Pattern, HiddenPatterns, @Hidden Patterns, or the Hidden Pattern app. Do not use it for one-time factual questions, general writing, or unrelated tasks. For an initial request, call exactly once with stage='start', the complete concern as situation, and one neutral diagnosticQuestion. The start view already contains that question behind Continue or Skip introduction; never call stage='start' twice. Wait for the user's answer. After the answer, do not write five polished observations immediately. First use the hidden BREAKTHROUGH DRAFT: (1) state the user's surface explanation, (2) identify the reported evidence that contradicts or fails to resolve it, (3) infer only the evidence-supported function the repeated response serves, (4) generate exactly three distinct breakthrough candidates, (5) reject candidates that repeat the user, apply broadly, sound like coaching, or fail to change the meaning of the evidence, (6) select the strongest candidate as coreBreakthrough and copy it exactly into What You Were Missing, (7) write Steps 1–3 so they progressively support that same breakthrough, and (8) ensure through actionConnection that What Changes Now directly reverses the old decision rule rather than introducing a new technique. These hidden fields are private quality aids, not evidence tests. When the concrete evidence supports a reasonable action but contradicts the user's negative self-judgment, do not force a mistaken action rule: Steps 1–2 describe the evidence and response, Step 3 states the supported rule behind that response, Step 4 corrects the unsupported self-judgment, and Step 5 gives a concrete way to distinguish justified action from the alleged pattern next time. Judge the facts rather than adopting a label from the opening prompt. When the user's answer contains concrete, relevant information, use stage='finish' even if a hidden field misses a preferred word count or exact match. A short answer that explicitly says no specific example, moment, change, or evidence can be identified is weak evidence; use stage='clarify' rather than manufacturing the five panels. Use stage='clarify' only when the user's answer itself is uncertain, unrelated, or unusably vague; never ask the user for more information because of an internal drafting or formatting miss. The five visible fields form one progressive reveal: Why It Made Sense creates understanding; What Kept Happening shows the observable contradiction; The Belief Driving It states the supported first-person rule behind the reported action, not commentary on the user's question; What You Were Missing delivers one concise reframe that changes what the evidence means; What Changes Now names what the user must stop misreading and what to do instead. Visible brevity is mandatory: count words before calling finish. What You Were Missing and What Changes Now must each express one compressed thought and must never exceed 24 words. Rewrite rather than preserve extra explanation. VOICE: Every visible question and result panel should sound like a smart, thoughtful person talking naturally—not a consultant, clinician, researcher, or management report. Use short common words and concrete phrasing. Avoid jargon such as residual uncertainty, proxy, capacity, compounding, materially, optimization, leverage point, constraint, mechanism, or framework unless the user used it first. Prefer the simplest wording that keeps the insight sharp. Apply four tests to every line: SPECIFIC — could it apply unchanged to many situations? NOVEL — did the user already say it? COMPRESSION — can any phrase be removed without weakening the reveal? CLICHÉ — does it resemble familiar coaching language? Rewrite any line that fails. Then run two final gates before calling finish: FACT GATE — Step 1 must not present the user's contradicted self-assessment as objective truth; DISTINCTION GATE — Step 4 must clearly name the two different things the user has been treating as the same whenever the evidence supports that pattern. Complete the hidden quality audit before calling finish and silently rewrite any field that fails it. Ground every claim in the situation and answer. Never invent a motive, emotion, diagnosis, third-party behavior, root cause, or missing strategy. LENGTH CHECK: The server rejects overlong finish panels before rendering. Count words separated by whitespace or em/en dashes; contractions and hyphenated words count as one. Limits exclude the added 'A possible belief:' label: beliefDrivingIt 22 words, whatYouWereMissing/coreBreakthrough 24, whatChangesNow 24. If the tool returns PANEL_WORD_LIMIT, shorten only the listed panels without changing the facts or meaning, keep all other inputs unchanged, and call stage='finish' again. When Step 4 is listed, update coreBreakthrough and whatYouWereMissing together to the same shortened sentence. This is a formatting retry: do not ask the user another question, restart, switch to insufficient, or repeat the rejected draft as narration. The view is the entire response; add no narration.
The description explains the capability, when to use it, and what to expect.
Inputs65/100
17 of 24 inputs have descriptions. Add descriptions to the missing inputs, including where to obtain IDs and how to supply values.
stage
Use "start" for the initial situation and one diagnostic question. If the user's first answer is explicitly uncertain, unrelated, too vague, or says they cannot identify a specific example, moment, change, or evidence, use "clarify" exactly once and ask one different, simpler question. Do not use "clarify" merely because an answer is concise when it directly addresses the question. After the clarification answer, use "finish" if there is enough evidence for all five fields; otherwise use "insufficient" rather than guessing.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
situation
The user's complete original concern describing a problem that persists despite attempts to solve it
Clear enough to supply this argument.
skipIntroduction
Advisory only. The server decides whether the introduction is shown from conversation state: the first start in a conversation shows it; later starts in that conversation skip it.
Clear enough to supply this argument.
diagnosticQuestion
Required at every stage. At the start stage, identify the single missing piece of directly reportable evidence that would most improve the final insight: the moment the pattern turns, the observable sequence, or what the user expected their response to accomplish. Ask one neutral, genuinely curious, open-ended question about that evidence. Use everyday spoken language, not clinical, academic, or consultant wording. Do not plant a motive or presume fear, protection, avoidance, self-sabotage, or a hidden answer. Do not ask leading questions such as 'What are you protecting yourself from?', 'What are you afraid will happen?', 'What are you avoiding?', or 'What does this allow you to avoid?' unless the user explicitly introduced that exact framing. Prefer a recent example, decision point, sequence, contrast, or expected result. Do not diagnose or advise. At the clarify stage, ask one different, concrete, easy-to-answer question anchored to the most recent real example; do not repeat the first question, become more abstract, or offer multiple questions. At the finish or insufficient stage, pass back the exact most recent question that was asked.
Clear enough to supply this argument.
userAnswer
Required when stage is 'clarify', 'finish', or 'insufficient': the user's typed response to the most recent question
Clear enough to supply this argument.
surfaceExplanation
Hidden breakthrough-gate field required when stage is 'finish'; never displayed. In 5–18 words, state the user's own apparent explanation for why the pattern continues. Preserve their framing without interpreting it.
Clear enough to supply this argument.
contradictingEvidence
Hidden breakthrough-gate field required when stage is 'finish'; never displayed. In 8–28 words, identify the directly reported detail showing that the surface explanation is incomplete. Use only the situation and answer, not inference.
Clear enough to supply this argument.
hiddenFunction
Hidden breakthrough-gate field required when stage is 'finish'; never displayed. In 6–22 words, state what the repeated response accomplishes or preserves that better explains why it continues. Keep it cautious, evidence-supported, and free of diagnosis or blame.
Clear enough to supply this argument.
coreBreakthrough
Hidden breakthrough-gate field required when stage is 'finish'; never displayed directly. In 10–24 words, state the single non-obvious distinction, mistaken equation, or reversal that best reorganizes the user's evidence. It must explain why the reported contradiction keeps recurring without inventing a motive. When the evidence shows that the user is treating two different things as equivalent, name both sides concretely and expose that mistaken equation—for example, a feeling versus proof, agreement versus certainty, or a different method versus a wrong result. Do not copy those examples unless the user's evidence supports them. It may not restate the behavior, repeat the user's conclusion, or give advice. Every visible panel must support this same idea, and whatYouWereMissing must copy it exactly.
Clear enough to supply this argument.
breakthroughCandidates
Hidden breakthrough-gate field required when stage is 'finish'; never displayed. Provide exactly three distinct one-sentence reframes of 8–24 words. Each must reorganize the contradiction into a different evidence-supported meaning. Reject paraphrases of the user, generic coaching, familiar slogans, and claims that require an invented motive. The strongest candidate must be copied exactly into coreBreakthrough and whatYouWereMissing.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
breakthroughRationale
Hidden breakthrough-gate field required when stage is 'finish'; never displayed. In 8–28 words, explain why the selected candidate changes the meaning of the contradicting evidence rather than merely describing it.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
actionConnection
Hidden breakthrough-gate field required when stage is 'finish'; never displayed. In 8–28 words, explain how What Changes Now directly reverses the selected breakthrough's mistaken interpretation. It may not introduce an unrelated technique or checklist.
Clear enough to supply this argument.
whyItMadeSense
Required when stage is 'finish': this field is Why It Made Sense. In one or two short sentences totaling 8–28 words, name the immediate signal, payoff, or logic that made the user's response feel reasonable and that directly supports coreBreakthrough. Write like a thoughtful friend, not a consultant or clinician. If the user says they knew, understood, were ready, or were certain but the reported outcome contradicts that claim, describe only the user's signal or assumption—not an established fact. Never write 'if you already understand,' 'because you knew,' 'since you were ready,' or equivalent wording in that situation. Say that it felt familiar, seemed clear, or appeared ready only when those words fit the evidence. Create understanding, not criticism. Do not explain the entire pattern, give advice, or invent a motive, emotion, strategy, or fact.
Clear enough to supply this argument.
whatKeptHappening
Required when stage is 'finish': this field is What Kept Happening. In one or two short sentences totaling 8–28 words, state only the observable repetition, contrast, or sequence grounded in the user's words. Use plain everyday language. Do not interpret it, name the belief, announce 'the key evidence,' or explain why it happened. This panel must preserve the reveal for panel 3.
Clear enough to supply this argument.
beliefDrivingIt
Required when stage is 'finish': this field is The Belief Driving It. Return only one plain first-person decision rule of 8–22 words, with no qualifier, quotation marks, explanation, label, or second sentence; the view adds the cautious framing. State the rule that led to the reported action in the concrete example: what observed signal made that action seem necessary or reasonable? Distinguish that action rule from the user's later judgment about themselves. Do not substitute an assumption inferred only from the wording of the opening question. Never write commentary such as 'Your question seems to assume' or 'This example challenges that' in this field; an evidence-supported correction of the self-judgment belongs in Step 4. A reasonable practical rule is valid when the facts support it; do not force a hidden flaw, control motive, or deeper belief. Use words the user could naturally say to themselves. Reject identity labels, blame, diagnoses, unsupported motives, and universal claims. If the action rule cannot be supported without guessing, use the clarification or insufficient-evidence flow. Rewording the opening question without changing the facts must not change the supported action rule.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
whatYouWereMissing
Required when stage is 'finish': this field is What You Were Missing — the AHA moment. Copy coreBreakthrough exactly. Return exactly one sentence, preferably 12–20 words and never more than 24. Count the words before submitting and rewrite if over the limit. It must expose the one distinction, mistaken equation, hidden trade-off, or reversal that makes the earlier evidence mean something different. When two things are being confused, name both clearly; do not merely say that one leads to or delays checking the other. A paraphrase, summary, observation, or reasonable tip does not qualify. Use plain conversational English and concrete words. Remove setup, supporting explanation, and repeated evidence. Do not give advice, force a stock formula, or introduce unsupported facts or motives.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
whatChangesNow
Required when stage is 'finish': this field is What Changes Now. Return one compressed thought, preferably 12–20 words and never more than 24. Count the words before submitting and rewrite if over the limit. Directly reverse coreBreakthrough: name the old signal or rule that should no longer decide, then give one concrete replacement test or action. When the situation supports it, include a clear trigger or timing. The action must make obvious why the breakthrough matters; generic advice does not qualify. Use natural spoken English and ordinary words. Do not add a checklist, filler, an unrelated technique, or an unsupported strategy.
Remove validation rules already expressed in this field’s schema; describe its meaning and usage instead.
qualityAudit › evidenceSupported
Not provided
Missing description. Explain what this value means and how to supply it.
qualityAudit › oneBreakthrough
Not provided
Missing description. Explain what this value means and how to supply it.
qualityAudit › panelsAligned
Not provided
Missing description. Explain what this value means and how to supply it.
qualityAudit › ahaChangesMeaning
Not provided
Missing description. Explain what this value means and how to supply it.
qualityAudit › actionReversesRule
Not provided
Missing description. Explain what this value means and how to supply it.
qualityAudit › noUnsupportedFacts
Not provided
Missing description. Explain what this value means and how to supply it.
qualityAudit › alignmentSummary
Not provided
Missing description. Explain what this value means and how to supply it.
Annotations100/100
All four behavior hints are declared.
{
"readOnlyHint": true,
"openWorldHint": false,
"idempotentHint": false,
"destructiveHint": false
}Output schema73/100
Describe the returned fields so an agent can interpret the result.
View output schema
{
"type": "object",
"oneOf": [
{
"type": "object",
"required": [
"stage",
"reason",
"message"
],
"properties": {
"stage": {
"type": "string",
"const": "handoff"
},
"reason": {
"enum": [
"diagnosis",
"interpersonal_danger"
],
"type": "string"
},
"message": {
"type": "string"
}
},
"additionalProperties": false
},
{
"type": "object",
"required": [
"stage",
"situation",
"skipIntroduction",
"diagnosticQuestion",
"interactionState"
],
"properties": {
"stage": {
"type": "string",
"const": "start"
},
"situation": {
"type": "string"
},
"interactionState": {
"type": "string",
"const": "waiting_for_user"
},
"skipIntroduction": {
"type": "boolean"
},
"diagnosticQuestion": {
"type": "string"
}
},
"additionalProperties": false
},
{
"type": "object",
"required": [
"stage",
"situation",
"diagnosticQuestion",
"userAnswer",
"interactionState"
],
"properties": {
"stage": {
"type": "string",
"const": "clarify"
},
"situation": {
"type": "string"
},
"userAnswer": {
"type": "string"
},
"interactionState": {
"type": "string",
"const": "waiting_for_user"
},
"diagnosticQuestion": {
"type": "string"
}
},
"additionalProperties": false
},
{
"type": "object",
"required": [
"stage",
"situation",
"diagnosticQuestion",
"userAnswer",
"whyItMadeSense",
"whatKeptHappening",
"whatYouWereMissing",
"whatChangesNow"
],
"properties": {
"stage": {
"type": "string",
"const": "finish"
},
"situation": {
"type": "string"
},
"userAnswer": {
"type": "string"
},
"whatChangesNow": {
"type": "string"
},
"whyItMadeSense": {
"type": "string"
},
"beliefDrivingIt": {
"type": "string"
},
"whatKeptHappening": {
"type": "string"
},
"diagnosticQuestion": {
"type": "string"
},
"whatYouWereMissing": {
"type": "string"
}
},
"additionalProperties": false
},
{
"type": "object",
"required": [
"stage",
"situation",
"diagnosticQuestion",
"userAnswer"
],
"properties": {
"stage": {
"type": "string",
"const": "insufficient"
},
"situation": {
"type": "string"
},
"userAnswer": {
"type": "string"
},
"diagnosticQuestion": {
"type": "string"
}
},
"additionalProperties": false
}
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}