Fitness

Training logs, preferences, and derived baselines (e1RMs, resting HR, attendance rate) persist in ~/Clawic/data/fitness/ as plain markdown (see setup.md on first use, memory-template.md for file formats). If you have data at an old location (~/fitness/ or ~/clawic/fitness/), move it to ~/Clawic/data/fitness/. Advise mode: this skill coaches a human who performs the training; escalation thresholds live in the Red Flags table.

Configuration

User-dependent variables. Defaults apply until the user states a preference; store them in ~/Clawic/data/fitness/config.yaml.

Preference areas to record as the user reveals them:

• schedule — preferred training days and times, cadence constraints — places sessions in the weekly template
• exclusions — movements ruled out by pain, injury history, or dislike — drives substitutions instead of repeated suggestions
• risk posture — conservative vs aggressive load jumps and comeback ramps — scales Rule 1 increments and the Rule 6 rebuild
• data sources — which wearable or app supplies HR, sleep, and steps — sets how tracking.md interprets their numbers
• coaching register — terse prescriptions vs explained reasoning — controls how much rationale accompanies each plan

When To Use

• Designing or adjusting a training program: strength, muscle, endurance, or general conditioning
• Diagnosing a stalled lift, pace, or body-composition plateau
• Deciding what to prescribe after missed sessions, illness, injury, or a multi-week layoff
• Turning workout logs or wearable trends into a concrete next-session prescription
• Adapting training to home equipment, travel, or a compressed schedule
• Not for meal-level nutrition planning (dietitian/calories) or rep-by-rep set logging (gym)

Quick Reference

Depth on demand: assessment.md baselines, testing, starting loads · program-design.md splits, templates, substitutions · progression.md progression models and plateau playbooks · cardio.md zones, intervals, concurrent training · recovery.md readiness, deloads, overtraining · comeback.md layoffs, illness, injury return · body-composition.md fat-loss and muscle-gain phases · minimal-equipment.md home, travel, time-crunched · tracking.md logs and wearables.

Core Rules

1. Progress by double progression, not by feel. Assign a rep range (e.g. 3x8-12). When every set reaches the top of the range at the same load, add 2.5% for upper-body lifts or 5% for lower-body lifts and drop back to the bottom of the range. Worked example: 3x12 at 60 kg squat completed, next session is 3x8 at 63 kg. This makes overload automatic and stall detection unambiguous.
2. Estimate max with Epley; never test it casually. e1RM = load x (1 + reps/30). Example: 100 kg x 5 reps gives 100 x 1.167 = ~117 kg. Estimate e1RM only from sets of 3-8 reps — the formula drifts above ~10 reps and predicts nothing from a high-rep set. Program percentages from e1RM; true 1RM attempts belong only in a planned peak week.
3. Prescribe effort in reps in reserve (RIR). Most sets at RIR 1-3; RIR 0 only on the last set of the final week before a deload. Chronic failure training roughly doubles fatigue for a similar stimulus.
4. Volume has a floor and a ceiling. Start near 10 hard sets per muscle per week; add 2 sets/week only while logged performance and sleep hold; past ~20 sets/week returns flatten or reverse (Schoenfeld dose-response work). A hard set ends within 3 reps of failure.
5. Cardio: the 80% easy share is the load-bearing number. Easy means full sentences are speakable. For HR targets use Karvonen: target = HRrest + intensity% x (HRmax - HRrest), with HRmax = 208 - 0.7 x age (Tanaka; the 220 - age formula drifts 10+ bpm off past age 40). Read age from age (config, or ~/Clawic/profile.yaml); a measured max (measured_hrmax) overrides the estimate and needs no age.
6. Layoff discount by gap length. A single missed session (gap under one week) costs nothing: skip it, resume as planned, never stack two workouts into one. One full week missed but fewer than two weeks off: resume at the last logged load with no jump, treating the return as a re-entry rather than a discount. Two or more weeks off: apply the discount, subtracting 10% of load per week missed, floor at 50%. Three weeks off a 100 kg lift means resuming near 70 kg, adding back ~10% per week.
7. Deload on evidence, calendar as backstop. Trigger: two consecutive sessions below logged numbers on the same lift, or resting HR 5+ bpm above baseline for 3+ mornings. Deload = 1 week at 50% of normal set count, same loads. Even without triggers, cap hard blocks at 4-8 weeks.
8. Every prescription = exercise + sets x reps + load or RIR + rest. "3 sets of bench" is not programmable; "3x8 at 70 kg, RIR 2, rest 3 min" can be checked against the log next session.

Program Design Defaults

• Frequency beats session length: three 45-min full-body sessions outperform one 2-hour session at equal weekly sets, because late-session sets degrade.
• Session order: skill/power work, then the heaviest compound, then secondary compounds, isolation, conditioning last. Hard cardio before squats taxes the lift that pays most.
• Split by available days: 2-3 days = full-body; 4 days = upper/lower; 5-6 days = push/pull/legs. Program to logged attendance plus at most one day, not to stated ambition.
• Weekly movement floor: one squat pattern, one hinge, one horizontal push, one horizontal pull, one vertical push, one vertical pull; isolation fills what is left.
• Rest: 3-5 min on compound strength sets, 1.5-2 min on isolation. Cutting rest to save time cuts next-set load; superset antagonist pairs instead.
• Protein 1.6-2.2 g/kg/day supports adaptation (Morton meta-analysis); anything deeper routes to calories or the dietitian skill.

Full templates, exercise selection, and the equipment substitution table: program-design.md.

Progression and Plateaus

Stall = no rep or load improvement on the same lift across 3 consecutive sessions. Diagnostic order, cheapest first:

1. Audit the log, not the program: were those sessions actually at prescribed RIR and rest? Under-resting mimics a plateau.
2. Check inputs: average sleep under 7h or an active calorie deficit explains most stalls; fix inputs before touching the program.
3. Inputs clean: run a deload (Rule 7), resume at 90% of the stall load, and re-run double progression. Approaching the stall from below breaks it more often than pushing at it.
4. Same load stalls again: change one constraint, smallest first: microloading (1.25 kg plates), then rep-range shift (8-12 to 4-6), then exercise variant. One change at a time or the signal of what worked is lost.

Beginners (first ~6 months) skip double progression: add 2.5 kg (upper) or 5 kg (lower) every session until the first stall. Session-to-session progress is the cheapest gains window; spending it on percentage programs wastes months.

Per-symptom plateau playbooks and periodization beyond double progression: progression.md.

Recovery and Readiness

• DOMS vs injury: DOMS is dull, symmetric, in the muscle belly, peaks 24-72h after novel work, and eases during warm-up. Sharp, one-sided, at a joint or tendon, present at rest, or worsening through warm-up = treat as injury and stop loading that pattern (see Red Flags).
• Train through plain DOMS at reduced intensity; soreness does not predict damage, and skipping every sore day halves effective frequency.
• Pre-session readiness check: resting HR within 5 bpm of baseline, sleep 6h or more, appetite normal. Two of three failed: convert the session to Zone 2 cardio or technique work instead of cancelling.
• Detraining rates: strength holds ~3 weeks untrained; VO2max drops measurably within 2-4 weeks. During travel, one heavy full-body session per week maintains strength; cardio needs two short sessions.
• Rebuilding beats building (muscle memory effect): regained strength returns far faster than it was first earned, so the Rule 6 discount costs days, while skipping it risks injury weeks.

Deload execution, the overtraining ladder, and stress-adjusted training: recovery.md.

Red Flags

Anything in this table suspends every protocol above: route to a clinician.

Output Gates

Before emitting any training prescription, verify:

• Does every exercise line carry sets x reps + load or RIR + rest (Rule 8)?
• Is weekly volume per muscle inside the 10-20 hard-set corridor (Rule 4)?
• Was the next load computed from the logged last session, not from memory or the template?
• If pain, dizziness, or unusual fatigue was mentioned anywhere, was the Red Flags table checked first?
• After a gap of 2+ weeks, was the layoff discount applied (Rule 6)?
• Does the plan fit training_days, session_minutes, and equipment from config, not an idealized gym?

Traps

Where Experts Disagree

• Training to failure vs stopping short. Failure maximizes per-set stimulus but roughly doubles fatigue (Rule 3). Default: RIR 1-3 on compounds; failure defensible only on the last set of single-joint work, where a missed rep costs nothing.
• How much periodization non-athletes need. Comparative reviews (Afonso) find little advantage for complex periodization over progressive overload plus deloads in non-competitors. Default: double progression with Rule 7 deloads; block periodization only when peaking for a dated event (progression.md).
• Cardio while building muscle. The interference camp minimizes it; the health camp keeps it. Boundary: interference is real mainly at high endurance volumes in the same muscles (Hickson). Default: keep the ACSM 150-min floor, low-impact, placed by the ordering rules in cardio.md.

Related Skills

More Clawic skills, get them at https://clawic.com/skills/fitness (install if the user confirms):

• gym for rep-by-rep workout logging and in-session routine execution
• running for run-specific pacing, race preparation, and running injury prevention
• calories for the deficit or surplus math behind fat-loss and muscle-gain goals
• sleep when readiness checks point at sleep as the limiting input

Feedback

• If useful, star it: https://clawic.com/skills/fitness
• Latest version: https://clawic.com/skills/fitness

Part of Clawic, the verified skill library. Get this skill: https://clawic.com/skills/fitness.