Conciseness

The body is dense with genuinely non-obvious operational detail (the FinnGen rsID exception, BBJ's GRCh37 build, the `pval: 0.0` underflow trap) and doesn't pad with concepts Claude already knows. It falls at level 4 rather than 5 because the ~200-word worked-example paragraph and some caveats repeated across the interpretation table and Limitations could be trimmed without losing operational value.

Actionability

Guidance is fully executable: copy-paste-ready calls with exact parameters (`UKBTOPMed_phewas_by_variant(rsid="rs7903146", max_pval=5e-8, limit=25)`, `Genebass_gene_burden_phewas(gene="PCSK9", burden_set="pLoF", max_pval=2.5e-6, limit=25)`), the required FinnGen `chr:pos:ref:alt` format with an example, and a fill-in report template. The worked example covers both common cases (variant-level and gene-burden).

Workflow Clarity

Five clearly sequenced steps with a decision point at Step 1 (variant-level vs gene-level), an interpretation table that functions as a checklist, and explicit verification loops such as "verify the ref/alt allele each biobank reported before interpreting" and "check `af` and `num_cases` before concluding biology". These feedback loops match the level-5 anchor; the operations are read-only lookups so no destructive-operation cap applies.

Progressive Disclosure

The file is well-sectioned with clean header navigation and no nested references, but it is ~96 lines with the worked example and the BBJ catalogue note inlined where they could live in a separate reference file. This matches the level-4 anchor (good structure, minor organization gaps); the level-5 anchor requires well-signaled one-level-deep references or a lean under-50-line overview.

Specificity

The description lists multiple concrete actions with named tools and biobanks: "look up every phenotype it associates with across European/UK (UKB-TOPMed), Finnish (FinnGen), Japanese (BioBank Japan), and Taiwanese (TPMI) biobanks, plus exome-wide gene-burden PheWAS (Genebass), then judge whether an association replicates across ancestries or is population-specific" and "compare effect sizes/allele frequencies of a variant across populations". This matches the comprehensive-coverage anchor; the level-4 anchor would require minor gaps in the action inventory, which are not evident.

Completeness

Both what (cross-biobank PheWAS lookup across four named ancestries plus a gene-burden layer, then replication judgment) and when ("Use whenever the user asks...") are explicit, with concrete trigger phrases plus explicit NOT-for boundaries. This is the level-5 pattern; level 4 would require a less explicit or specific 'when' clause, which is not the case here.

Trigger Term Quality

It quotes natural user phrases verbatim — "what else is this variant/gene associated with", "does this association replicate in other ancestries / biobanks", "is this effect East-Asian-specific", "pleiotropy of rsXXX", "phenome scan" — covering synonyms and paraphrases users would actually say, which exceeds the good-coverage-but-few-missing level 4 anchor.

Distinctiveness Conflict Risk

The description explicitly disambiguates adjacent skills — "NOT for the forward direction (trait → which SNPs: use the gwas-* skills), NOT for fine-mapping a locus (use tooluniverse-gwas-finemapping), and NOT for single-SNP mechanism tracing in one population (use tooluniverse-gwas-snp-interpretation)" — giving it a clear niche with minimal conflict risk, matching the level-5 anchor rather than the minor-overlap level 4 anchor.