Skill Distiller

Compress verbose skills to reduce context window usage. This skill is self-compressed using formula notation (~400 tokens, ~90% functionality, LLM-estimated). Full reference version: SKILL.reference.md.

Note: This skill uses formula notation — the LLM executes these operations directly. You don't need to understand the math. For prose explanation, see SKILL.reference.md.

Legend

S = {TRIGGER, CORE, CONSTRAINT, OUTPUT, EXAMPLE, EDGE, EXPLAIN, VERBOSE}
I(s) ∈ [0,1]        # importance score
P = {yaml.name, yaml.desc, N-count, task-create, checkpoint, BEFORE/AFTER}  # protected
θ ∈ [0,1]           # threshold (default 0.9)
n ∈ ℕ               # target tokens

Operations