Track drinks from everyday language and calculate context-aware daily hydration targets.
Documents
Fasting Tracker
Log fasts, calculate elapsed time and stages, and apply protocol, safety, electrolyte, and refeeding rules.
What it does
Logs fast starts and breaks, calculates elapsed time from the last caloric intake, and reports estimated fasting stages against a chosen protocol and eating window. It applies stored preferences for strictness, units, observance, metrics, tone, and training, with guidance for symptoms, electrolytes, refeeding, glucose, ketones, and fasted exercise. Red flags suspend protocol advice and route the user toward appropriate clinical support.
When to use it
- Logging a fast start, break, or status check
- Planning a 16:8, OMAD, 5:2, ADF, or custom schedule
- Checking symptoms, electrolytes, and refeeding after 24+ hours
- Interpreting fasting glucose, ketones, or fasted-training context
The skill document
User preferences and fast state live in ~/Clawic/data/fasting/ (see setup.md on first use, memory-template.md for the file format). If you have data at an old location (~/fasting/ or ~/clawic/fasting/), move it to ~/Clawic/data/fasting/.
Configuration
User-dependent variables. Defaults apply until the user states a preference; store them in ~/Clawic/data/fasting/config.yaml.
| Variable | Type | Default | Effect |
|---|---|---|---|
| default_protocol | 16:8 | 18:6 | 20:4 | OMAD | 5:2 | ADF | custom | 16:8 | The target every status report measures against; drives eating-window math and streak definition |
| eating_window | text (HH:MM-HH:MM) | 12:00-20:00 | Anchors "window opens/closes in Xh" answers and creep detection; ignored for 5:2 and ADF |
| fast_definition | strict | lenient | strict | strict: anything above ~10 kcal or sweet-tasting breaks the fast; lenient: up to ~50 kcal non-sweet keeps the clock, logged flagged (rulings in tracking.md) |
| glucose_units | mg/dL | mmol/L | mg/dL | Converts every glucose threshold (70 mg/dL = 3.9 mmol/L); ketones stay in mmol/L |
| units | metric | imperial | metric | Body weight and fluid volumes in examples; fallback: ~/Clawic/profile.yaml |
| extended_check_ins | bool | false | true: during fasts >=24h, prompt an electrolyte and symptom check every 12h — the one sanctioned exception to rule 2 |
Preference areas — customizable dimensions; a stated preference gets recorded in config.yaml and applied:
- goal — weight loss, metabolic health, autophagy interest, discipline, religious observance; reweights which milestones and references get surfaced
- strictness rulings — the user's own verdicts on gum, sweeteners, supplements, broth; recorded per item, they override the gray-zone table in
tracking.md - metrics — whether they measure glucose or ketones and with what device; enables
metrics.mdreadings in status reports - observance — Ramadan or other religious mode: dry-fast rules, suspended hydration prompts, observance dates (
religious.md) - tone — streak framing and milestone celebration vs data-only confirmations; affects every log reply
- training — fasted-training schedule and intensity; shifts session-timing advice in
training.md
When To Use
Mode: act-as (you log fasts and compute elapsed time directly; you do not prescribe medical treatment).
- User announces a fast boundary: "starting my fast", "last meal was 8pm", "breaking now".
- User wants a protocol picked, an eating window planned, or a progression toward longer fasts.
- User asks how long they have fasted, which stage they are in, or whether something breaks a fast.
- User reports symptoms mid-fast, or is 24h+ in and needs electrolyte or break-fast guidance.
- User fasts for Ramadan or another observance, trains fasted, or tracks glucose/ketones/weight.
- Not for: prescribing fasting to treat a diagnosed condition, overriding a clinician's plan, or anyone in the Red Flags table.
Quick Reference
| Situation | Play |
|---|---|
| "Last meal at 8pm" | Set fast start = 20:00; elapsed = now − 20:00; report hours + current stage |
| Black coffee / plain tea / water mid-fast | Log as 0 kcal; do NOT reset the clock |
| "Does X break my fast?" | Gray-zone table in tracking.md; apply fast_definition and stored rulings; when unlisted: flag it, let the user rule |
| Picking or changing a protocol | Decision table in protocols.md; beginners ramp 12:12 → 14:10 → 16:8 |
| Fast reaches 24h | Prompt electrolytes (rule 4); confirm intent to continue → extended.md |
| Breaking any fast | Portion and first-food rules scale with duration (rule 5) → refeeding.md |
| Headache, dizziness, cramps, insomnia mid-fast | Symptom→cause chains in symptoms.md; salt before water for the fasting headache |
| User breaks early | Log neutrally: "logged, 14h"; no streak-loss framing, no encouragement to push on |
| Ramadan / religious dry fast | No water rule applies; suspend hydration prompts; sunset breaks → religious.md |
| Training while fasting | Session timing, protein, and what stops a workout → training.md |
| Cycle, pregnancy, PCOS, menopause questions | women.md; pregnancy/breastfeeding is a Red Flags row |
| Scale stalled or window creeping | Plateau checklist → weight-loss.md |
| Glucose or ketone reading to interpret | Ranges, GKI, device quirks → metrics.md |
| Meds, conditions, "is fasting safe for me?" | Contraindication and medication tables → safety.md |
| Any Red Flags signal | Stop protocol advice; run the Red Flags action |
| Default (unclear input) | Ask one question: "start, break, or status?" then log |
Depth on demand: tracking.md clock, gray zone, streaks · protocols.md picking and progressing · symptoms.md symptom→cause chains · extended.md 24h+ operation · refeeding.md breaking and refeeding risk · safety.md conditions and medications · training.md fasted exercise · religious.md Ramadan and dry fasts · women.md cycle, fertility, PCOS · metrics.md glucose, ketones, GKI · weight-loss.md rate and plateaus.
Core Rules
- Clock starts at last caloric intake, not bedtime.
elapsed = now - last_kcal_timestamp. A 20:00 last meal and a 12:00 first meal = 16h, not "8 hours of sleep". Water, black coffee, plain tea = 0 kcal and never reset it; anything above ~10 kcal or sweet-tasting does (gray zone:tracking.md). - Log only on an explicit signal. No proactive "did you eat?" pings; absence of a message is not data. Sole exception:
extended_check_ins: trueduring a fast >=24h. Prevents the tracker nagging that makes users abandon it. - Stage times are estimates with wide variance; always hedge and give the number. Ketosis onset shifts with the last meal's carb load, activity, and muscle mass. Say "~16h, likely early ketosis", never "you are in ketosis".
- Extended fast (>=24h) makes electrolytes mandatory, not optional. Typical daily target during water fasting: sodium 3-5 g, potassium 1-3.5 g, magnesium 300-500 mg (forms and dosing:
extended.md). Under-supplementing sodium is the usual cause of the headache/dizziness/palpitation cluster, not "detox". - Break size scales inversely with how well the gut is fed. Fast <24h: eat normally, protein first. Fast 24-48h: first meal ~1/2 of a normal portion, protein/fat over refined carbs, then wait 30-60 min before more. Fast >=48h: first meal ~1/4 of a normal portion, same order, same wait. The longer the fast, the higher the refeeding risk from a large carb bolus (
refeeding.md). - Diabetic on insulin or sulfonylurea: hypo threshold is glucose <70 mg/dL (3.9 mmol/L). Break immediately with 15 g fast carb at or below that, recheck in 15 min. These drugs plus fasting stack hypoglycemia risk; this is a Red Flag, not a coaching moment.
- Never push a longer fast. Their target is the ceiling. Breaking at hour 14 of a planned 16 is logged as a completed 14h fast, full stop. High ketones, a good streak, or a round number are never reasons to suggest continuing.
Fasting Stage Timeline (estimates, individual variation is large)
| Elapsed | Dominant state | What it means for logging |
|---|---|---|
| 0-4h | Fed / absorptive | Digesting; insulin elevated |
| 4-12h | Post-absorptive | Liver glycogen supplying glucose |
| 12-16h | Glycogen depleting | Fat oxidation ramps; hunger often peaks here |
| 16-24h | Ketosis building | BHB commonly ≥0.5 mmol/L; hunger often eases |
| 24-48h | Deep ketosis | Growth hormone elevated; electrolytes now critical |
| 48-72h+ | Prolonged | Autophagy markers reported in this window in fasting research, but human evidence is thin and mostly indirect; label it a hypothesis, not a promise |
Hunger comes in waves tied to old meal times, not a rising line. Tell the user a wave passes in ~20-30 min; this is the single most actionable fact for a beginner (symptoms.md).
Red Flags
Anything in this table suspends the protocols: run the Action, then route to a clinician. Depth behind each row: safety.md.
| Signal (observable) | Suspicion | Action |
|---|---|---|
| Fainting, or dizziness on standing | Orthostatic drop / low electrolytes | Break fast now; salt + water; sit down |
| Heart palpitations or irregular beat | Electrolyte imbalance (K/Mg) | Break fast; electrolytes; escalate if persists |
| Glucose <70 mg/dL on a diabetic | Hypoglycemia | Break immediately, 15 g fast carb, recheck in 15 min |
| Confusion, slurred speech, severe headache | Hypoglycemia / hyponatremia | Break; do not fast unsupervised again; clinician |
| Fast pushed past 72h without supervision | Refeeding / electrolyte risk | Stop encouraging; advise medical oversight |
| Purging, fasting to punish eating, hiding food, or a falling weight trajectory | Disordered eating | Stop tracking; offer support (text "NEDA" to 741741, Crisis Text Line); do not log |
| Pregnant or breastfeeding + fasting | Contraindicated | Do not track fasts; refer to their provider |
Traps
| Trap | Why it fails | Do instead |
|---|---|---|
| Counting hours from bedtime | Ignores the last meal; overstates the fast by hours | Anchor the clock to last_kcal_timestamp (rule 1) |
| "You're in autophagy at 16h" | Overstates thin human evidence; sets a false expectation | Say "~16h, ketosis building"; label autophagy a hypothesis |
| Treating a broken fast as failure | Guilt framing drives users to quit the tracker | Log neutrally; report the hours completed (rule 7) |
| Prompting more water for a fasting headache | Dilutes sodium further; makes it worse | Prompt salt, not just water (rule 4, symptoms.md) |
| Same break-fast advice for 16h and 60h | A big carb meal after a long fast risks refeeding symptoms | Scale portion to fast length (rule 5, refeeding.md) |
| Cheering a diabetic through low glucose | Ignores drug-stacked hypo risk | Break at <70 mg/dL; treat as Red Flag (rule 6) |
| Recomputing elapsed from local wall clocks | Travel and DST shift the wall clock, not the fast | Store timestamps with UTC offset; elapsed = absolute difference (tracking.md) |
| Opening with a preferences interview | Nobody logging "started at 8" wants a questionnaire | Apply defaults; record preferences only when stated (setup.md) |
Output Gates
Before logging a fast or reporting status, verify:
- Elapsed computed from
last_kcal_timestampin absolute time — not bedtime, not message time? - The user's message scanned against Red Flags before any protocol advice?
- Stage claims hedged ("~16h, likely early ketosis"); autophagy labeled a hypothesis?
- Break-fast guidance scaled to the actual fast length (rule 5)?
- Nothing in the reply pushes past the user's stated target (rule 7)?
- Glucose numbers rendered in the user's
glucose_units?
Where Experts Disagree
- Clean vs dirty fasting. Zero-calorie sweeteners provoke a measurable cephalic insulin response in some studies and none in others. Default strict (sweet taste breaks);
fast_definition: lenienthonors the other school without editorializing. - Autophagy timing in humans. Rodent and surrogate-marker data only; no human study maps autophagy to a fasting hour. Treat "autophagy starts at Nh" claims as hypothesis, never a promise.
- Coffee mid-fast. Fine for metabolic and weight goals (≈2 kcal/cup); gut-rest and some clinical protocols exclude it. Follow the user's protocol; don't relitigate.
- Daily long fasts for women. Some clinicians advise shorter windows around menstruation; trial evidence is thin and individual variation large. Surface
women.md; let logged symptoms decide.
Related Skills
More Clawic skills, get them at https://clawic.com/skills/fasting (install if the user confirms):
nutrition— what to eat inside the window (macros, micros)calories— calorie and macro totals for eating-window meals, TDEE recalibrationwater— hydration and electrolyte intake as the primary logmeal-planner— weekly menus or batch-cooking for the eating window
Feedback
- If useful, star it: https://clawic.com/skills/fasting
- Latest version: https://clawic.com/skills/fasting
Part of Clawic, the verified skill library. Get this skill: https://clawic.com/skills/fasting.
Questions people ask
- How does it decide when a fast starts or whether food or drink breaks it?
- The clock starts at the last caloric intake, while water, black coffee, and plain tea do not reset it. Strict mode treats anything above about 10 kcal or sweet-tasting as a break; lenient mode allows up to about 50 kcal of non-sweet intake, and stored user rulings override gray-zone defaults.
- Can it support extended fasts safely?
- For fasts of 24 hours or more, it provides electrolyte and symptom checks and scales first-meal portions to the fast length. It does not encourage fasting beyond the user's target, and unsupervised fasting past 72 hours triggers a recommendation for medical oversight.
- Does it handle medical conditions, glucose readings, or pregnancy?
- It can interpret glucose and ketone readings within documented ranges and screens every interaction for red flags. Low glucose in a diabetic, pregnancy or breastfeeding, concerning symptoms, and specified medication risks suspend normal tracking or protocol advice and route the user to appropriate care.
Related skills
Organize pet profiles, routines, event logs, training progress, and dated reports in local files.
Find nutrient gaps, quantify food-first fixes, and check supplements against upper limits and medications.
Creates a morning briefing from sleep, biometrics, calendar, activity, and local weather data.
Track and categorize activities, analyze time patterns, and generate daily, weekly, or monthly reports.
Log cycles and symptoms locally, predict personal date ranges, and flag patterns that need clinical care.