Conciseness

The body is dense and expert-level — motif grammar diagram, parameter tables, published hit rates, exact thresholds — with no padding explaining concepts Claude already knows. It is not 5 because the Workflow Overview block partially duplicates the per-phase sections, the licence paragraph is long, and the "LOOK UP, DON'T GUESS" / "COMPUTE, DON'T DESCRIBE" boilerplate adds tokens without skill-specific payoff; it is well above 3 since nearly all content is non-obvious and specific.

Actionability

Concrete, executable-level guidance throughout: script names (`run_generator.py`, `evaluate_design.py`), exact defaults (`num_sampling_steps: 1000`, LigandMPNN "temperature 0.1 with weights `v_32_030_25`"), a fully specified `motif_str` grammar with diagram, and MPD parameters ("re-noise the design by 575 of 1000 steps… require 9/10 to pass"). It is not 5 because no runnable artifact is shown — the manifest JSON that drives `run_pipeline.py` is never exemplified, and exact tool-call syntax for generation/resequencing is absent; it is not 3 because the parameter-level specificity far exceeds "some concrete guidance".

Workflow Clarity

Eight phases are clearly sequenced with an explicit pre-GPU validation checkpoint (Phase 2: "A malformed `motif_str` caught here costs seconds; caught in Phase 6 it costs the campaign"), an explicit error-recovery loop ("If a campaign fails, re-examine the motif before re-tuning filters"), inter-phase dependency notes (Phase 5 must be self-arranged; Phases 6–7 depend on it), and a mandated cumulative funnel report. This matches the 5 anchor — clear sequence, explicit validation, feedback loops for a batch operation — rather than the 4 anchor's "most checkpoints present".

Progressive Disclosure

References are one level deep and clearly signaled ("See `FILTERING.md`", "See `CAVEATS.md`", plus a dedicated Reference files section), but the referenced files do not exist in the bundle — there is no `references/` directory, and the out-of-bundle paths (`skills/setup-apnovo-remote-tool/SKILL.md`, `src/tooluniverse/remote/apnovo/README.md`) are likewise unverifiable. Broken reference targets make navigation fail as shipped, which is worse than the 4 anchor's "minor organization gaps", while the clear signaling and appropriate inline/split judgment keep it above the 3 anchor's "buried references".

Specificity

The description lists multiple concrete actions covering the full pipeline — "build a transition-state motif, generate scaffolds with AlphaProtein Novo, fold every mechanistic state, and triage with mechanism-inspired filters" — with the filter detail ("catalytic geometry, multi-state H-bonds, multi-resequence backbone scoring") going beyond generic claims. Coverage of what the skill does is comprehensive rather than having minor gaps, so it matches the 5 anchor rather than 4.

Completeness

Both questions are answered explicitly: the "what" is the four concrete pipeline actions, and the "when" is a dedicated trigger clause — "Use for designing enzymes for a named reaction, scoring or filtering existing designs, or deciding which designs to order". This matches the 5 anchor's "clearly and explicitly answers both what AND when with concrete trigger phrases"; the when-clause is explicit, not merely present or weakly implied, so it is above 4.

Trigger Term Quality

Natural phrasings users would say are present — "designing enzymes for a named reaction", "scoring or filtering existing designs", "deciding which designs to order" — but common synonyms are missing: "biocatalyst", "enzyme engineering", and "optimize an existing design" (which the body's When-to-Use covers but the description omits). Good keyword coverage with a few natural terms missing fits the 4 anchor; it is not 5 because synonym coverage is not comprehensive.

Distinctiveness Conflict Risk

"De novo enzyme design by catalytic-motif scaffolding" plus a named generator carves a clear niche with distinct triggers, and the skill body explicitly routes adjacent requests (binder design, structure prediction alone) to other skills. Minimal conflict risk with any generic skill matches the 5 anchor.