Lean and table-driven with no concept-explaining filler Claude already knows (no 'a PDF is...' style padding); every section earns its place, including the gotchas and honest-limitations notes.

Provides a copy-paste-ready executable command ('tu run DrugSynergy_calculate_bliss ...') with expected output, named tool calls, and per-tool parameter lists in the Step 0 table, plus a real referenced script with concrete CLI args.

Clear Step 0→1→2→3→4 sequence with explicit validation checkpoints (consistent-scale warning, ceiling-effect headroom check, 'state which model you used', Hill-fit data requirements) and an error-recovery-aware gotchas section.

Organized overview with sectioned steps and tables; the single referenced bundle file scripts/synergy_reference.py is real and one level deep, with no nested reference chains; cross-skill links live in a clearly labeled 'Related skills' section.

Names a concrete action ('quantify whether two drugs together are synergistic, additive, or antagonistic') and enumerates five specific reference models (Bliss, HSA, Loewe, ZIP, Chou-Talalay CI), matching the 'multiple specific concrete actions' anchor.

Answers both what (quantify synergy/additivity/antagonism via named models) and when ('Use when you have measured single-drug and combination effects (inhibition/viability) and need a synergy score'), satisfying the explicit-trigger requirement.

Uses natural terms a user would say — 'synergy', 'drug combination', 'synergistic, additive, or antagonistic', 'inhibition/viability', 'synergy score' — giving good coverage rather than only jargon or a single keyword.

Clear niche (reference-model synergy scoring) plus an explicit exclusion clause ('NOT for looking up pre-computed synergy in a database (use the SYNERGxDB tool / cell-line-profiling skill)') that distinguishes it from a sibling skill.