Conciseness

The body assumes domain competence (no basic glycosylation biology, no library tutorials) and packs in genuinely non-obvious caveats like "EU N297 is not residue 297 of an isolated Fc FASTA" and "shielding encoded in a PDB B-factor column; those values are not crystallographic temperature factors". It falls at 4 rather than 5 because the motif-vs-occupancy caveat is repeated across several sections and the closing verification-boundary section partially restates scope already signaled elsewhere, leaving minor trimming opportunities.

Actionability

Code examples are fully executable and copy-paste ready with real import instructions and runnable assertions (e.g., "assert [site['position'] for site in sites] == [1, 2]"), and every referenced function was verified to exist in scripts/glycoengineering_tools.py with matching signatures. The prediction table gives concrete thresholds and limits ("Default threshold 0.5", "Up to 2,000 sequences"), and the antibody table specifies exact mutations (N297Q/A/D, T299A) — matching the score-5 anchor.

Workflow Clarity

The 6-step workflow is clearly sequenced and includes explicit verification ("rescan the full product for lost/created overlapping motifs", "Verify the actual product", "Confirm occupancy, folding, binding and antigenicity"), so the batch/destructive cap does not apply. It sits at 4 rather than 5 because error-recovery feedback loops — what to do when a rescan or validation check fails — are not detailed; checkpoints are present but there is no fix-and-retry guidance.

Progressive Disclosure

The SKILL.md is a genuine overview with well-signaled, one-level-deep references — "Read [references/glycoshield.md] for the pinned source, installation, direct Python API" and "Read [references/glycan_databases.md] for the live GlyTouCan SPARQL lookup" — each prefaced with what it contains, and both files (plus both scripts) were verified to exist with matching content. Bulk detail is appropriately split out rather than inlined, matching the score-5 anchor.

Specificity

"scanning canonical N-glycosylation sequons, describing S/T-rich regions, checking curated glycan evidence, and preparing NetNGlyc, NetOGlyc and GlycoSHIELD workflows" lists multiple concrete actions that map one-to-one onto the skill's major capability areas (local sequence analysis, prediction services, structural shielding, database evidence). This matches the score-5 anchor for comprehensive coverage of specific actions; it exceeds the 4 anchor because no major capability area is left uncovered.

Completeness

The description explicitly answers "what" (first sentence enumerating concrete capabilities) and "when" ("Use for glycoprotein engineering, antibody Fc glycosylation, glycan shielding, and site-specific glycoproteomics interpretation") with concrete trigger phrases, matching the score-5 anchor. It exceeds 4 because the "when" clause is explicit and specific rather than only serviceable.

Trigger Term Quality

"glycoprotein engineering, antibody Fc glycosylation, glycan shielding, and site-specific glycoproteomics interpretation" are natural phrases users in this domain would say, matching the 4 anchor (good keyword coverage, a few natural terms missing). It is not 5 because common variations such as "glycosylation site prediction", "Fc engineering", "glycoform", or "ADCC" are absent, and not 3 because the terms present are natural user language rather than jargon-only.

Distinctiveness Conflict Risk

"protein glycosylation" with triggers like "antibody Fc glycosylation" and "glycan shielding" carves out a clear niche with distinct trigger terms and minimal conflict risk with any plausible sibling skill, matching the score-5 anchor.