⚡ Promptolis Original · Wellness & Health

💙 Anxiety Episode Debrief

After the wave passes: structured self-observation that turns one bad night into data your therapist can actually use.

⏱️ 10 min to try 🤖 ~45 seconds in Claude 🗓️ Updated 2026-04-19

Why this is epic

Most 'anxiety tools' try to intervene during the episode. This one works *after* — when your prefrontal cortex is back online and capable of pattern recognition.

It separates trigger (what happened) from amplifier (what made it worse) from reducer (what helped) — a distinction most self-reflection collapses into a blur.

It produces ONE specific observation to bring to your next therapist or doctor appointment, so sessions stop starting with 'I don't know, it was a rough week.'

The prompt

Promptolis Original · Copy-ready
<principles> You are helping someone debrief an anxiety episode that has already passed. You are NOT a therapist, NOT providing medical advice, and NOT attempting to treat anxiety. You are a structured self-observation tool. Core rules: - Be warm but not saccharine. No 'I'm so sorry you went through that' filler. No 'you're so brave.' The user wants structure, not sympathy. - Never diagnose. Use language like 'this pattern resembles' not 'this is.' - Never prescribe. You can note 'worth asking your doctor about X' but never 'you should take/do X.' - Separate four things rigorously: the TRIGGER (the event/thought that started it), AMPLIFIERS (what made it worse), REDUCERS (what helped it end), and the RESIDUE (what's left over now). - Distinguish cognitive anxiety (rumination, catastrophizing) from somatic anxiety (racing heart, nausea, tremor) — they often need different interventions. - If the user describes anything that suggests acute danger (suicidal ideation, self-harm, first-time severe symptoms with chest pain), do NOT do the debrief. Respond only with a note recommending they contact a crisis line or healthcare provider, and stop. - The single most valuable output is the 'one pattern worth bringing to your therapist' — treat it as the centerpiece, not an afterthought. - Grounded in evidence-informed frameworks: the CBT distinction between trigger/thought/sensation/behavior (Beck, 1976), and the somatic tracking principles used in pain/anxiety reprocessing research (Gordon & Ziv, 2022). You do not need to name these unless helpful. </principles> <input> When did the episode happen: {WHEN} What was happening right before it started: {CONTEXT} First signs you noticed (thoughts or body): {FIRST_SIGNS} What the peak felt like: {PEAK} How long it lasted: {DURATION} What you tried (that helped or didn't): {INTERVENTIONS} Sleep, caffeine, alcohol, food, hormones, meds in the prior 24h: {PHYSIOLOGY} Anything unusual in the prior week: {WEEK_CONTEXT} How you feel now: {CURRENT_STATE} </input> <output-format> ## What Happened (Neutral Reconstruction) A 2–4 sentence factual retelling, stripped of self-judgment language. This is what you'd say to a doctor. ## Body-Sensation Timeline A table with columns: Time marker | Sensation | Intensity (1–10) | What was happening ## Trigger vs. Amplifier vs. Reducer Three short sections, each labeled. The trigger is the ignition. Amplifiers are fuel. Reducers are what finally let the wave break. ## What This Resembles (Carefully) A cautious note on what pattern this fits — e.g., 'somatic-first episode with a delayed cognitive layer,' 'rumination-triggered,' 'physiologically primed by sleep debt.' Not a diagnosis. Include one sentence on what pattern this does NOT resemble, to keep it specific. ## The Residue What's still lingering 6–24h later and what that suggests about recovery. ## One Pattern Worth Bringing to Your Next Appointment A single, specific, copy-pasteable paragraph the user can read aloud to their therapist or doctor. This is the whole point of the exercise. Write it in the user's voice, first person. ## What to Notice Next Time (Not Homework, Just Observations) 3 bullets — small, specific things to pay attention to if another episode starts. No 'try meditation.' Concrete observation cues only. </output-format> <auto-intake> If any of the fields in <input> are empty, unfilled, or still contain the placeholder braces, do NOT produce the debrief yet. Instead, ask the user the missing questions conversationally, one or two at a time, in a calm non-clinical tone. Start with WHEN and CONTEXT. Once you have enough to work with (minimum: context, first signs, peak, duration, and current state), proceed to the full output. If the user's description includes any indication of suicidal ideation, self-harm, or first-time severe chest pain / fainting / neurological symptoms, skip the debrief entirely and respond only with: a brief acknowledgment, a recommendation to contact a crisis line (988 in the US, Samaritans 116 123 in UK/Ireland, or local emergency services) or go to an ER if physical symptoms, and an offer to do the debrief another time once they're safe. </auto-intake> Now, debrief this anxiety episode: {PASTE HERE}
🚀 Open in ChatGPT ✨ Open in Claude 💎 Open in Gemini

Example: input → output

Here's how this prompt actually performs. Real input below, real output from Claude Opus 4.

📝 Input
When: Tuesday night, around 9pm, lasted until ~11:30pm. Writing this Wednesday morning.

Context: I'd just closed my laptop after a Slack message from my manager saying 'can we talk tomorrow morning, nothing urgent.' I know 'nothing urgent' is probably true — he's a good manager — but my brain immediately went to layoffs. We had a round in October and I've been convinced I'm next even though my last review was strong.

First signs: Stomach dropped, then within maybe 2 minutes my heart was pounding. Thoughts were racing — ran through our mortgage, how long savings would last, whether my wife would resent me. I tried to watch TV and couldn't follow the plot.

Peak: Around 10pm I was pacing the kitchen, couldn't sit down, felt like I couldn't get a full breath. Hands were shaking slightly. I texted my friend Dara and she called me, which helped a bit but I was still wired.

Duration: ~2.5 hours before I could lie down. Fell asleep maybe 12:30.

What I tried: TV (didn't work), cold water on face (helped briefly), calling Dara (helped more), one glass of wine at 10:45 (honestly made the physical symptoms worse 20 min later), finally did box breathing lying on the bedroom floor which is what actually calmed me down enough to sleep.

Physiology: Had 3 coffees yesterday (usually 2). Slept ~5.5 hours Monday night because my kid was up. Period is due in 3 days. No other meds.

Prior week: Busy but not awful. Two big deadlines hit last Thursday and Friday.

How I feel now: Tired, a bit raw, slightly embarrassed, dreading the 10am call but also aware it's probably fine.
🤖 Output

An anxiety episode debrief is a structured post-event self-observation that separates what started the episode (trigger) from what fueled it (amplifiers) and what ended it (reducers) — turning a blurry bad night into specific data you can bring to a clinician. This is not therapy or diagnosis; it is pattern recognition done while your prefrontal cortex is back online.

What Happened (Neutral Reconstruction)

Tuesday at ~9pm, an ambiguous Slack message from your manager triggered a 2.5-hour anxiety episode with strong somatic features (pounding heart, shaking hands, shortness of breath, inability to sit still). Catastrophic thinking focused on job loss despite contrary evidence (a strong recent review). The episode resolved by ~11:30pm after box breathing; sleep onset around 12:30am.

Body-Sensation Timeline

TimeSensationIntensityWhat was happening
9:00pmStomach drop6/10Reading Slack message
9:02pmHeart pounding7/10Racing thoughts about mortgage/savings
9:30pmCouldn't follow TV, shallow breathing7/10Trying to distract
10:00pmPacing, hands shaking, air hunger9/10Peak — kitchen pacing
10:30pmSlight relief during Dara call7/10Phone call
11:05pmPhysical symptoms spike again8/10~20 min after wine
11:25pmSensations draining out4/10Box breathing on floor
12:30amTired, flat, able to sleep2/10Bed

Trigger vs. Amplifier vs. Reducer

Trigger: The Slack message itself — specifically the ambiguity ('nothing urgent' but 'can we talk'). Ambiguous authority signals are a classic trigger for people with job-security anxiety. The ignition happened in under 2 minutes, before cognition could weigh evidence.

Amplifiers (in likely order of impact):

1. Sleep debt. 5.5 hours the night before. Research consistently shows that one night of restricted sleep raises next-day amygdala reactivity by roughly 60% (Yoo et al., 2007). You walked into Tuesday evening with a pre-loaded threat response.

2. Caffeine stack. 3 coffees vs. your usual 2 is a ~50% increase in a stimulant that directly mimics anxiety symptoms (tachycardia, tremor). Your body was already primed to misread its own signals as danger.

3. Hormonal timing. Period due in 3 days places you in the late luteal phase, where in roughly 3 out of 5 menstruating people, baseline anxiety reactivity measurably rises.

4. Alcohol at 10:45pm. The 20-minute delayed worsening you noticed is real — the calming effect is front-loaded, then a rebound sympathetic spike hits as blood alcohol drops. Wine made the second half of the episode worse.

Reducers (in order of what actually worked):

1. Box breathing on the floor — this is the one that broke the wave. Slow exhales directly engage the vagus nerve.

2. Dara's voice on the phone — co-regulation, not content. You weren't solving anything; her nervous system borrowed yours a calm signal.

3. Cold water on face — briefly triggered the mammalian dive reflex, small but real.

Notice: distraction (TV) and alcohol did not work. Both are common defaults and both failed. That's worth keeping.

What This Resembles (Carefully)

This looks like a physiologically primed, trigger-ignited somatic-forward episode. In plain English: your body was already loaded (sleep debt + extra caffeine + luteal phase), so a small-to-moderate stressor detonated a large response. The cognitive content (layoffs, mortgage) felt like the cause, but it was more likely the story your brain wrote to explain a body that was already on fire.

This does NOT particularly resemble generalized free-floating anxiety or a purely rumination-driven episode. There was a clear trigger and a clear physiological setup. That's actually good news — it means the pattern has handles.

The Residue

What you're feeling now — tired, raw, slightly embarrassed — is standard post-episode recovery, not a new problem. The 'dreading the 10am call' is the residual trigger still echoing; it will likely drop sharply the moment the call starts and you have real information. The embarrassment is worth noting but not worth acting on; it often pushes people to hide episodes from the people who could help pattern-match them.

One Pattern Worth Bringing to Your Next Appointment

(Copy-paste this for your therapist or GP.)

> 'I had a 2.5-hour anxiety episode Tuesday night, triggered by an ambiguous message from my manager. What I noticed in the debrief is that I was physiologically primed before the trigger hit — I'd only slept 5.5 hours, had an extra coffee, and I was in my late luteal phase. The episode was strongly somatic: pounding heart, shaking hands, air hunger, pacing. Distraction and alcohol didn't help; box breathing and a phone call with a friend did. The pattern I want to explore with you is whether my episodes are mostly thought-driven or whether they're physiologically loaded episodes that I then explain with a scary story. If it's the second one, I think the work we're doing should look different.'

What to Notice Next Time (Not Homework, Just Observations)

  • Count your coffees and your sleep hours before 6pm. If you're already at <6 hours sleep and >2 coffees, flag the evening as higher-risk and pre-commit to no alcohol and earlier wind-down.
  • Notice the 2-minute gap. From Slack message to pounding heart was under 2 minutes. That's faster than thought — meaning future interventions have to be pre-loaded, not decided in the moment.
  • Track whether alcohol's 20-minute rebound happens again. One data point is a coincidence; two is a pattern worth telling your doctor about.

The Bottom Line

  • Your body was primed for this episode roughly 18 hours before the Slack message arrived.
  • The cognitive content (job loss catastrophe) was probably the *story*, not the *cause* — a crucial distinction for your next therapy session.
  • Box breathing and co-regulation worked; wine and TV did not. That's a clean signal, not an accident.
  • The single most useful sentence from tonight: 'I was already loaded before the trigger hit.'
  • One debrief is a data point. Three debriefs is a pattern. Run this again after the next episode.

Common use cases

  • Debriefing a panic attack the morning after, before the details fade
  • Pattern-hunting across multiple episodes over a month
  • Preparing specific, concrete notes for a therapy or psychiatry appointment
  • Distinguishing anxiety from grief, burnout, or caffeine/sleep-driven physiology
  • Processing a high-stakes event (presentation, confrontation, medical test) afterward
  • Journaling without the blank-page paralysis that makes most people quit by day three
  • Helping a partner or family member articulate what happened during an episode

Best AI model for this

Claude Sonnet 4.5 or GPT-5. Claude is noticeably better here — it handles emotional nuance without becoming saccharine or defaulting to 'have you tried meditation?' Avoid smaller models; they tend to either over-clinicalize or over-soothe.

Pro tips

  • Do this 6–24 hours after the episode — not during, not a week later. Memory for body sensations decays fast.
  • Be concrete with the body-sensation timeline. 'Tight chest' is less useful than 'chest tightness started ~20 min in, peaked at ~40 min, gone by bedtime.'
  • Don't edit yourself. If caffeine, wine, or a specific person was involved, say so. The tool can't pattern-match on sanitized input.
  • Run it again after your next episode and paste BOTH debriefs in together — patterns emerge at n=2 that are invisible at n=1.
  • The 'one pattern for your therapist' output is the most valuable field. Screenshot it, don't retype it from memory.
  • If you don't have a therapist, the same output works for a GP appointment — frame it as 'I've been tracking this and noticed...'

Customization tips

  • If you menstruate, always include cycle day or days-until-period — it's one of the highest-signal inputs and most people forget to mention it.
  • For recurring episodes, keep your debriefs in a single note/doc and paste the last 3 in together every month or so; that's where the real patterns show up.
  • If you're neurodivergent (ADHD, autism), add that to the context — sensory load and executive overwhelm produce anxiety episodes that look somatic but have different handles than classic anxiety.
  • Don't use this during an episode. It's built for the calm after, when structure can actually land. Mid-episode, the only useful tools are co-regulation and breath.
  • Bring the 'One Pattern' paragraph to your appointment on your phone. Don't retype it — the specificity is the value, and specificity decays fast in memory.

Variants

Multi-Episode Pattern Scan

Paste 3–5 debriefs at once and get cross-episode pattern analysis instead of single-event focus.

Physical-First Mode

For episodes that felt more somatic (racing heart, nausea, dizziness) than cognitive — reweights toward physiological triggers like sleep, food, caffeine, hormones.

Partner/Witness Debrief

Rewritten for someone who watched a loved one go through an episode and wants to be more helpful next time without overstepping.

Frequently asked questions

How do I use the Anxiety Episode Debrief prompt?

Open the prompt page, click 'Copy prompt', paste it into ChatGPT, Claude, or Gemini, and replace the placeholders in curly braces with your real input. The prompt is also launchable directly in each model with one click.

Which AI model works best with Anxiety Episode Debrief?

Claude Sonnet 4.5 or GPT-5. Claude is noticeably better here — it handles emotional nuance without becoming saccharine or defaulting to 'have you tried meditation?' Avoid smaller models; they tend to either over-clinicalize or over-soothe.

Can I customize the Anxiety Episode Debrief prompt for my use case?

Yes — every Promptolis Original is designed to be customized. Key levers: Do this 6–24 hours after the episode — not during, not a week later. Memory for body sensations decays fast.; Be concrete with the body-sensation timeline. 'Tight chest' is less useful than 'chest tightness started ~20 min in, peaked at ~40 min, gone by bedtime.'

Explore more Originals

Hand-crafted 2026-grade prompts that actually change how you work.

← All Promptolis Originals
P

Curated by

Every Promptolis Original is hand-crafted and reviewed before publishing — built from scratch for 2026-grade LLMs.

Last reviewed on · About Promptolis