Score the Process, Pace It to the Outcome
The skill now works the way the plan was actually built: outcomes are read, leading indicators are scored, and habits are paced and layered so keeping them gets you there.
Why Care?
The skill said "score inputs, not outcomes" but never said what to do with the outcome someone actually came for. Told "I want to lose a lot of weight", it had no move. And it never pushed back on the most common way these plans fail: starting too many habits at once.
What's New?
references/goal-pacing.md, the philosophy as a method. Process goals are leading indicators, scored weekly. Outcomes are lagging indicators, paced to milestones at the program form's own Day 30, 60, and 90 checkpoints and scored only there, from a Day 0 baseline.Bring an outcome: Claude makes it concrete, finds the levers, and paces them with real rates (and web research where it matters) so the outcome is plausible by day 100, showing the math.
Bring a process goal: Claude projects a reasonable outcome, as a range, split across the Day 30, 60, and 90 milestones.
Layer, don't stack: start below what feels like enough, add one habit or one step of dose at a time, add up the daily load across all goals, deload in Weeks 8 and 12.
Safe numbers: ranges over false precision, conservative health rates (sustained fat loss runs roughly 0.5–2 lb a week; early low-carb drops are mostly water), doctor sign-off for fasting and hard training.
The skill classifies every goal as outcome or process before drafting it, writes the weekly plan as a ramp, checks total load before the closing sections, and scores outcomes against their milestones at Day 30, 60, and 90.
The scored template gains an outcome line with its Day 0 baseline, a pacing section, outcome milestones in the program's By 30 / 60 / 90 Days table, and a milestone-vs-actual table.
The splash and README say the same thing in plain words.