Conciseness

Mostly dense operational content (tool calls with literal parameters), but the `EBIProteins_get_rna_editing` broken-endpoint warning is repeated three times (Phase 4d, Fallbacks, Limitations), the coordinate-mapping tools are covered twice with overlapping prose, and multiple date-stamped status notes ('checked 2026-10-08', 'verified live') are not confined to a deprecated/status section. Not score 4 because the repetition and duplicated EBI prose go beyond minor trimming.

Actionability

Calls are copy-paste ready with real parameter values — `iPTMnet_get_ptm_sites(operation="get_ptm_sites", uniprot_id="P04637")`, `EBIProteins_get_proteins_by_genomic_loc(taxonomy="9606", location="17:7676154")` — and include invocation syntax (`tu run EBIProteins_get_rna_editing '{...}'`), verified outputs, and expected-result examples (TP53 mapping, insulin cleavage). Not score 4 because the common cases are covered with executable, literal examples.

Workflow Clarity

The Phase 0→5 sequence is clear with per-phase fallbacks and some checkpoints (probe iPTMnet once for 503 before switching, 'always check total_entries/total_variants' for isoform ambiguity, evidence-graded synthesis). Not score 5 because Phase 5 'Synthesis & Report' is named in the workflow but has no corresponding section defining what the report must contain.

Progressive Disclosure

Section structure and tables are good, but there is no bundle: the ~50 lines of EBI Proteins tool tables, the verified-live examples, and the Tool Parameter Reference are exactly the API-reference material that belongs in a separate reference file inlined into SKILL.md. The single external pointer (`src/tooluniverse/data/ebi_proteins_ext_tools.json`) is a repo path, not a clearly-signaled one-level-deep bundle reference. Not score 4 because a substantial, separable reference block is inline.

Specificity

The description lists multiple concrete actions — 'PTM site annotation, kinase-substrate identification, and PTM-disease associations' — and enumerates specific tools with their roles ('iPTMnet (sites + enzymes), ProtVar (functional consequences), UniProt (baseline)'). Coverage is comprehensive for the domain; not score 4 because there are no gaps in the action inventory.

Completeness

Both 'what' (PTM analysis using the named tools, with the EBI Proteins API 'as a cross-validation source') and 'when' ('Use for PTM site annotation, kinase-substrate identification, and PTM-disease associations') are explicitly and concretely stated. Not score 4 because the when-clause is explicit and trigger-phrased rather than merely implied.

Trigger Term Quality

Full-term-plus-abbreviation ('Post-translational modification (PTM)') and all five major modification types ('phosphorylation, ubiquitination, acetylation, glycosylation, methylation') plus 'kinase-substrate' and 'PTM-disease' give comprehensive natural-keyword coverage. Not score 4 because users' natural phrasings for this domain are essentially all present.

Distinctiveness Conflict Risk

The PTM-analysis niche is clear with distinct triggers, but the description also claims 'protein-level variant lookup by accession/dbSNP/HGVS', which overlaps with a sibling variant-analysis skill for rsID/HGVS queries. Not score 5 because that variant-lookup mention creates minor wrong-skill trigger risk; not score 3 because the overlap is confined to one sub-capability.