Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Two linked protocols for the two most common mental health problems in post-white-collar survival: burnout and broken sleep. They feed each other — burnout destroys sleep, broken sleep accelerates burnout. Fix them together.
.claude/skills/mental-reset-suite/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | — | — |
| case-04 | ✗→✓ | ▲ Improved | — | — |
| case-21 | ✗→✓ | ▲ Improved | — | — |
| case-09 | ✗→✓ | ▲ Improved | — | — |
| case-20 | ✗→✓ | ▲ Improved | — | — |
Two linked protocols for the two most common mental health problems in post-white-collar survival: burnout and broken sleep. They feed each other — burnout destroys sleep, broken sleep accelerates burnout. Fix them together.
Module A is the flagship burnout protocol based on the Maslach Burnout Inventory framework. Module B is the evidence-based sleep overhaul. Both include agent-driven assessments, tracking, and automation triggers.
Burnout:
Sleep:
Module A (Burnout) — use when:
Module B (Sleep) — use when:
Both together — use when:
Agent action: Administer the assessment interactively — ask each question one at a time, record the score, and calculate totals. Store the scores and burnout stage in agent state.
BURNOUT ASSESSMENT (answer 1=never to 5=every day)
EXHAUSTION:
[ ] I feel emotionally drained by my work
[ ] I feel used up at the end of the day
[ ] I dread getting up to face another day
[ ] Working all day is really a strain
CYNICISM:
[ ] I've become more callous toward people since this job
[ ] I don't really care what happens to some people I work with
[ ] I feel like I'm just going through the motions
[ ] I doubt the significance of my work
INEFFICACY:
[ ] I can't deal with problems effectively anymore
[ ] I feel I'm not making a difference
[ ] I've accomplished less than I used to
[ ] I don't feel excited about my achievements
SCORING:
- 12-24: Mild burnout — prevention stage. Focus on boundaries.
- 25-40: Moderate burnout — active recovery needed.
- 41-60: Severe burnout — consider medical leave, therapy, or major life change.Do NOT try to fix everything at once. This week is about stopping the bleeding.
Agent action: Set daily evening reminders for the screen cutoff time. Create a "Week 1 Priorities" note with only the user's top 3 must-do items. Schedule a check-in for end of Week 1.
Sleep protocol:
The "absolute minimum" exercise:
One recovery activity per day:
Agent action: Help the user customize each script below with their specific names, times, and situations. Save customized scripts to ~/documents/burnout-recovery/boundary-scripts.txt. Track which boundaries have been set and enforced.
BOUNDARY SCRIPTS — copy and use as needed
DECLINING EXTRA WORK:
"I want to do a good job on what I already have. If I take this on,
something else will suffer. Can we talk about what to deprioritize?"
PROTECTING OFF-HOURS:
"I'm offline after [time]. If it's genuinely urgent, text me.
Otherwise I'll see it tomorrow morning."
SAYING NO TO MEETINGS:
"Can I get the summary notes instead? I want to protect my focus
time so I can deliver [specific thing] on time."
PUSHING BACK ON SCOPE:
"I can do A or B this week, not both. Which one matters more?"Track every activity for one work week. Mark each as:
Agent action: Send a daily prompt at end of workday asking the user to rate their activities. Compile the week's data. Identify the top 3 drains and top 3 energizers.
ENERGY AUDIT TEMPLATE
PATTERNS TO IDENTIFY:
-> Which tasks drain you most? Can any be delegated?
-> Which people drain you most? Can you reduce contact?
-> When is your best energy? Protect those hours.
-> What energizes you that you've stopped doing?Agent action: Guide the user through these reflection questions over several sessions. Record their answers and surface them during low moments.
Ask yourself:
The goal isn't to love every minute. It's to have enough meaning to offset the hard parts.
After 4-6 weeks of active recovery, ask honestly:
If no on 2+ of those: this isn't a recovery problem, it's an environment problem. Start planning your exit. See the layoff-72-hours and career-reinvention skills.
Agent action: Ask the user each question below, one at a time. Record answers in agent state. Use the results to generate a personalized protocol.
SLEEP AUDIT
1. What time do you currently go to bed? (average)
2. What time do you currently wake up? (average on weekdays)
3. What time do you actually fall asleep after getting into bed?
4. How many times do you wake up during the night?
5. How do you feel within 30 minutes of waking? (1 = exhausted, 5 = alert)
6. Do you use alcohol to help you sleep?
7. Do you use screens (phone, TV, laptop) in bed?
8. What time is your last caffeine most days?
9. Do you nap? If yes, how long and when?
10. Is your bedroom dark? Cool? Quiet?
RED FLAGS:
- Falls asleep in under 5 min: may be chronically sleep-deprived
- Wakes unrefreshed regardless of duration: possible sleep apnea
(recommend doctor consult)
- Uses alcohol to sleep: address this first
- Last caffeine after 2pm: address in Step 4The single most powerful intervention is a consistent wake time — 7 days a week, including weekends.
Agent action: Ask "What time must you be awake for work or responsibilities?" Set a daily morning reminder at that time. Track wake time adherence in state.
SLEEP ANCHOR RULES:
1. Pick one wake time. Keep it every day, including weekends.
Variation of more than 30 minutes breaks the system.
2. Do NOT "catch up" on sleep on weekends.
Weekend sleep debt recovery resets your anchor.
3. Get out of bed within 10 minutes of your alarm.
4. Your body will correct bedtime naturally within 1-2 weeks
once the anchor is consistent.
WHY IT WORKS:
Sleep drive (adenosine) builds from the moment you wake.
Consistent wake time = consistent sleep drive at the same hour
each night = falling asleep faster and sleeping deeper.Agent action: Help the user build their specific ritual from the options below. Set a daily reminder 35 minutes before their target bedtime.
WIND-DOWN RITUAL BUILDER — pick 2-3 from each tier
START (35 min before bed):
[ ] Dim all lights or switch to lamps only
[ ] Stop eating — digestion competes with sleep
[ ] Set out tomorrow's clothes, bag, keys
[ ] Write tomorrow's 3-item priority list
MIDDLE (20 min before bed):
[ ] Take a warm shower or bath (the body temperature drop
afterward triggers sleep onset)
[ ] Read a physical book or e-ink reader (no backlit screens)
[ ] Do 5 minutes of gentle stretching
[ ] Write a "brain dump" — everything in your head, on paper
FINAL (10 min before bed):
[ ] Phone face-down on charger outside the bedroom
[ ] Drop bedroom temperature to 65-68F / 18-20C if possible
[ ] 4-7-8 breathing: inhale 4s, hold 7s, exhale 8s (x3)
[ ] No clock-checking — turn clocks awayCaffeine has a half-life of 5-7 hours. To have less than 25% caffeine remaining at bedtime, your last caffeine = bedtime minus 12 hours.
CAFFEINE CUTOFF EXAMPLES:
Bedtime 10:30pm → last caffeine by 10:30am
Bedtime 11:30pm → last caffeine by 11:30am
Bedtime midnight → last caffeine by noon
HIDDEN CAFFEINE SOURCES:
- Green tea: 25-45mg per cup
- Dark chocolate: 12-25mg per ounce
- "Decaf" coffee: 2-25mg per cup
- Pre-workout and energy drinks: read the label
- Some pain relievers (Excedrin): 65mg per tabletSLEEP ENVIRONMENT CHECKLIST:
DARKNESS (most impactful after timing):
[ ] Blackout curtains or sleep mask
[ ] No LED indicator lights in bedroom (cover with black tape)
[ ] No screen glow at all
TEMPERATURE:
[ ] Bedroom at 65-68F / 18-20C
[ ] Fan, lighter duvet, cotton sheets, open window if you run hot
QUIET (or consistent noise):
[ ] White noise machine or fan
[ ] Earplugs if partner snores
STIMULUS CONTROL (bed is for sleep only):
[ ] No working in bed
[ ] No TV in bed
[ ] No scrolling in bed
[ ] If you can't sleep after 20 minutes: GET UP. Do something
boring in dim light until sleepy. Return to bed only when sleepy.
(This is counterintuitive. It works.)Agent action: After 14 days, re-administer the sleep audit. Compare to baseline.
IF NO IMPROVEMENT after 14 days with anchor + environment fixes:
Consider: sleep apnea screening (home sleep tests often covered
by insurance), or referral to sleep medicine specialist for CBT-I
(Cognitive Behavioral Therapy for Insomnia — more effective than
any sleep medication, no side effects).
CBT-I providers: behavioralsleep.orgBurnout:
Sleep:
Persist across sessions:
yamlmental_reset: burnout: assessment_scores: exhaustion: null cynicism: null inefficacy: null total: null date: null assessment_history: [] burnout_stage: null current_week: 0 recovery_start_date: null boundaries_set: [] boundaries_enforced: [] energy_audit: energizers: [] drains: [] completed_days: 0 meaning_reflections: [] exit_decision: null sleep: audit: bedtime_current: null wake_time_current: null sleep_onset_minutes: null night_wakings: null morning_feel_score: null uses_alcohol: null uses_screens_in_bed: null last_caffeine_time: null audit_date: null sleep_anchor_time: null anchor_adherence_days: 0 wind_down_ritual: [] caffeine_cutoff_time: null protocol_start_date: null checkin_scores: [] flags: possible_sleep_apnea: false alcohol_dependency: false anxiety_driven: false
yamltriggers: - name: burnout_evening_wind_down condition: "burnout.current_week <= 4" schedule: "daily at user's chosen cutoff time" action: "Gentle reminder: screens off soon. How was today on a 1-5 scale? Log any recovery activity." - name: burnout_energy_audit_prompt condition: "burnout.current_week >= 3 AND burnout.current_week <= 4 AND burnout.energy_audit.completed_days < 5" schedule: "daily at end of workday" action: "End of day energy check: What did you do today? Rate each activity E/N/D." - name: burnout_weekly_checkin condition: "burnout.recovery_start_date IS SET" schedule: "weekly on Sunday evening" action: "Weekly recovery check-in: How are you feeling compared to last week? Review boundaries set vs enforced." - name: burnout_reassessment condition: "burnout.current_week >= 6" schedule: "once" action: "Time for reassessment. Re-administer the burnout assessment and compare to your initial baseline. If scores haven't improved, recommend professional support." - name: sleep_morning_anchor condition: "sleep.sleep_anchor_time IS SET" schedule: "daily at sleep.sleep_anchor_time" action: "Time to get up. Keeping this consistent is the most important thing you can do for your sleep. How did last night go? (1-5)" - name: sleep_wind_down_start condition: "sleep.wind_down_ritual IS SET" schedule: "daily 35 minutes before target bedtime" action: "Wind-down time. Dim your lights now. Screens off in 25 minutes." - name: sleep_two_week_checkin condition: "sleep.protocol_start_date IS SET" schedule: "14 days after sleep.protocol_start_date" action: "Two-week sleep check-in. Let's re-run the sleep audit and compare to your baseline. Ready?" - name: depression_flag condition: "burnout.assessment_scores.total >= 50 OR (6 weeks elapsed AND scores not improving)" action: "Burnout may coexist with depression. Recommend speaking with a mental health professional. Crisis line: call or text 988."
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-05 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-08 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-12 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-22 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-07 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-04 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-14 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-11 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-17 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-02 | pass→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-03 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-21 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-09 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-20 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-18 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-19 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-01 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-06 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-13 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-15 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-16 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +23 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.
Other measured skills in the registry, with their headline benchmark lift.