CBT CardsPatterns
Name a recurring thinking–behavior loop without turning it into a diagnosis.
CBT Cards / Content library
More than cards: four ways to make an idea usable.
24 CBT Cards-owned modules: patterns, bounded experiments, worked examples, and reflection scripts. Each points back to the existing reviewed practice system instead of creating a second recommendation layer.
Four formats, four jobs
Recognize. Test. See. Phrase.
Experiments
Compare a low-stakes prediction with an observable result instead of arguing with yourself.
Worked examples
Show what a structured breakdown looks like in an ordinary situation.
Scripts
Use sentence starters as scaffolding, not affirmations or belief instructions.
How this fits with cards and metaphors
Cards are compact prompts. Metaphors make an idea easier to remember. Reviewed practices connect a plain-language situation to one bounded action. This library adds four editorial forms around that trusted layer: recognize a loop, test a low-risk prediction, inspect a worked example, or borrow a precise sentence structure.
This is CBT Cards-owned content, deliberately separate from the pinned MetalHatsCats source toolkit. It does not change publication status for any raw source card, metaphor, or protocol.
Machine-readable version: data/content-library.json, including stable record IDs, linked practice IDs, evidence IDs, and the full experiment safety exclusions.
Safety boundary
General wellness and self-reflection only. Do not use these modules to override genuine danger, protective boundaries, required safety/professional/accessibility rules, or medical, legal, financial, or safeguarding decisions. If a module conflicts with its linked reviewed practice, the practice safety boundary controls.
6 modules
Patterns
Recognize the loop before trying to change it.
Prediction as Fact
A feared outcome starts being treated as if it has already been observed.
Useful question: What do I know directly, and what am I predicting?
Related: Test Your Prediction
The One-More-Check Loop
One extra check briefly reduces uncertainty, then teaches the mind to ask for another.
Useful question: Is this check adding information, or only trying to change how certain I feel?
Related: One Less Check
Replay Without Learning
An event is reviewed repeatedly after the useful facts, lesson, or next action have stopped changing.
Useful question: What new fact, decision, repair, or action did the last round produce?
Related: Reflection or Replay?
The Task Gets Bigger While You Wait
Avoiding a safe task makes the imagined task larger, more complete, and harder to start.
Useful question: What is the smallest safe part I could approach now without making completion the goal?
Related: The Smallest Approach Step
The Moving Finish Line
A task has no stable definition of done, so each completed pass creates another reason to keep polishing.
Useful question: What would count as done if I had to define it before the next pass?
Related: Set the Good-Enough Contract
From One Event to an Identity Verdict
A specific mistake, awkward moment, or weak result is converted into a global statement about who you are.
Useful question: What exactly happened, in behavior-level language, without turning it into a verdict about the whole person?
6 modules
Experiments
Turn a low-stakes prediction into a bounded observation.
One Normal Reread
Test whether an ordinary low-stakes message really needs repeated rereading before it can be sent.
Prediction: If I stop after one normal review, I will miss something important or regret sending it.
Bounded test: Use a genuinely low-stakes message. Read it once for meaning and obvious errors, then send it without restarting the review cycle.
Update: What did the result show about the difference between useful checking and certainty-seeking?
Only for ordinary, reversible, low-stakes situations. Preserve required safety/professional/accessibility rules and avoid high-stakes decisions. Related: One Less Check, Test Your Prediction
Ask Without Full Rehearsal
Test a small social prediction by asking one ordinary question without scripting every possible branch first.
Prediction: If I do not fully rehearse, I will freeze, sound unacceptable, or be unable to recover.
Bounded test: Choose an ordinary, low-risk question. Prepare the first sentence only, ask it, and allow the rest of the exchange to unfold normally.
Update: Which part of the prediction was observed, and which part remained a prediction?
Only for ordinary, reversible, low-stakes situations. Preserve required safety/professional/accessibility rules and avoid high-stakes decisions. Related: Test Your Prediction, Spot the Safety Behavior
Five-Minute Start
Test whether starting a safe avoided task requires feeling ready first.
Prediction: If I start before I feel ready, the task will be unmanageable and I will get stuck.
Bounded test: Choose one reversible five-minute step that approaches the task without requiring completion.
Update: What did direct contact teach that planning from a distance could not?
Only for ordinary, reversible, low-stakes situations. Preserve required safety/professional/accessibility rules and avoid high-stakes decisions. Related: The Smallest Approach Step
Stop at Written Criteria
Test whether a discretionary task can stop at a prewritten fit-for-purpose standard instead of at a feeling of complete certainty.
Prediction: If I stop when the written criteria are met, the result will be obviously inadequate.
Bounded test: Before another pass, write three concrete completion criteria for an ordinary discretionary task. Stop when all three are met unless genuinely new information appears.
Update: Would you change the criteria next time based on evidence, or only to remove discomfort?
Only for ordinary, reversible, low-stakes situations. Preserve required safety/professional/accessibility rules and avoid high-stakes decisions. Related: Set the Good-Enough Contract
Two Dimensions, No Verdict
Test what becomes visible when a result is rated on specific dimensions instead of passed through one success/failure label.
Prediction: If one important part went badly, the whole result was a failure.
Bounded test: Choose an ordinary completed task. Rate two concrete dimensions from 0–10, such as clarity and timeliness, without assigning one total verdict.
Update: What action follows from the specific ratings that did not follow from “failure”?
Only for ordinary, reversible, low-stakes situations. Preserve required safety/professional/accessibility rules and avoid high-stakes decisions. Related: Use a Continuum, Not a Verdict
Same-Standard Swap
Test whether a self-rule remains the same when only the person being judged changes.
Prediction: My harsher rule is simply the objectively correct standard.
Bounded test: Describe the same ordinary situation twice: once with you in it and once with another person who has genuinely similar duties, information, and constraints. Write the rule for each.
Update: What is the narrowest rule that respects real responsibility without adding a person-specific penalty?
Only for ordinary, reversible, low-stakes situations. Preserve required safety/professional/accessibility rules and avoid high-stakes decisions. Related: Run the Same-Standard Check
6 modules
Worked examples
See how observation, inference, standards, and next actions can be separated.
No Reply Yet
Situation: You sent an ordinary message this morning and have not received a reply by lunchtime.
Before: They have not replied. They must be annoyed, and I probably said something wrong.
Breakdown: Observation: There is no reply yet. Interpretation: They may be annoyed with me. Unknown: The reason for the delay is not known. Useful next step: Unless the message is time-sensitive, wait for more information instead of treating the interpretation as an observation.
The revised version does not force a positive story. It separates what is known from what is inferred.
Related: Test Your Prediction
A Mistake Without an Identity Sentence
Situation: During a presentation, you give one number incorrectly and correct it a minute later.
Before: I am incompetent.
Breakdown: Specific behavior: I gave one number incorrectly and corrected it. Responsibility: I should check why I missed it and whether the correction reached the audience. Identity boundary: The event does not by itself establish a global claim about my competence.
The specific description keeps accountability while leaving something concrete to review.
Problem or Hypothetical Worry?
Situation: You keep thinking, “What if I forget something important for tomorrow?”
Before: I need to keep going through every possible thing I might forget.
Breakdown: Practical problem: There is a short list of items that genuinely need to be ready tomorrow. Action: Write the list once and prepare the items that can be prepared now. Remaining worry: The possibility that there is some unknown thing you have not remembered cannot be solved by endlessly repeating the scan.
The split protects the practical action instead of dismissing the whole concern as “just worry.”
Related: Problem or Worry?, Park and Return
A Report With a Finish Line
Situation: A short internal report is complete, but you keep changing wording because it still does not feel finished.
Before: I should keep improving it until nothing bothers me.
Breakdown: Criteria: The requested questions are answered.; Figures match the source data.; The main recommendation is understandable to the intended reader. Stopping rule: If these criteria are met and no new factual issue appears, discomfort about stopping is not automatically a new requirement.
The standard is defined before another pass and can be revised later from actual feedback rather than from endless hypothetical defects.
Related: Set the Good-Enough Contract
One Awkward Moment on a Continuum
Situation: A conversation had one awkward pause, but you also asked the question you needed to ask and received a useful answer.
Before: That conversation was a failure.
Breakdown: Clarity: 7/10 — the question was understandable. Smoothness: 5/10 — there was an awkward pause. Usefulness: 8/10 — the needed information was obtained. Next step: If anything needs work, practice a clearer opening rather than treating the whole interaction as failed.
Specific dimensions can preserve a real weakness without erasing the rest of the outcome.
Related: Use a Continuum, Not a Verdict
From Replay to Review
Situation: After a meeting, you have replayed the same comment several times.
Before: I need to go through it again until I understand exactly why I said it that way.
Breakdown: New information: No new facts appeared in the last two replays. Lesson: Next time, pause before answering a question you did not expect. Repair or action: No repair is currently needed; if new information appears, review then. Stop point: End this round because the review has produced a lesson and no further action.
Stopping is tied to whether the review is still producing information, not to whether the event feels emotionally complete.
Related: Reflection or Replay?
6 modules
Scripts
Borrow sentence structure without being told what conclusion to believe.
Observation / Inference
A thought contains both something directly observed and a conclusion about what it means.
Template: What I know directly is ____. The interpretation I am adding is ____. What would give me more information is ____.
Guardrail: Do not use this wording to dismiss credible warning signs or to force an uncertain situation to sound harmless.
Related: Test Your Prediction
Action / Worry
A concern keeps returning and you are unsure whether more thinking is producing a practical step.
Template: The concrete problem, if any, is ____. The action I can take now is ____. The part I cannot solve right now is ____, so I will review it again at ____.
Guardrail: If delaying action could worsen a real problem, take or seek the appropriate action instead of parking it.
Related: Problem or Worry?, Park and Return
Enough Information Without Total Certainty
An ordinary low-risk decision has enough information but still does not feel certain.
Template: I would prefer certainty about ____. This decision actually requires ____. The information I already have is ____. One more check would add ____ / would mainly reduce discomfort.
Guardrail: Do not apply this to required checks, high-stakes decisions, or situations where material information is genuinely missing.
Related: One Less Check
Good-Enough Contract
A discretionary task keeps expanding because “done” has not been defined.
Template: For this task, fit for purpose means: 1) ____, 2) ____, 3) ____. I will stop when these are met unless new factual information changes the requirement.
Guardrail: This does not lower legitimate safety, accuracy, compliance, or professional standards.
Related: Set the Good-Enough Contract
Behavior, Not Identity
A specific mistake is being turned into a global self-label.
Template: The specific behavior was ____. Its effect was ____. What I am responsible for now is ____. The global label I am tempted to add is ____, but that label is broader than the event.
Guardrail: Specific language should preserve responsibility and repair where they are actually needed.
Replay Exit
You have already reviewed an event and are unsure whether another round is useful.
Template: The new information from the last round is ____. The decision, lesson, or repair is ____. If nothing new is appearing, my next action is ____ and this review ends here for now.
Guardrail: Do not stop a review that is still needed for a concrete safety issue, accountability step, or decision with genuinely missing information.
Related: Reflection or Replay?
Editorial method
These formats are editorial syntheses mapped to existing CBT Cards practice mechanisms and evidence records. That mapping keeps terminology, exclusions, and provenance consistent; it does not establish independent clinical efficacy for the exact pattern, experiment, example, or script format.
For recommendation decisions, use the reviewed Practice Finder and preserve its avoid_when rules.