Conciseness

The body is mostly efficient but has clear tightening opportunities: dosage-sensitivity thresholds appear twice (HI/TS score rules in 'SV Pathogenicity Reasoning' and again in 'Phase 3'), every tool is listed per-phase and then re-listed verbatim in 'Required Tools Reference', and the 'KEY PRINCIPLES' list restates points already made in the reasoning section. Not 4: the duplication is more than minor — a reader processes the same HI/TS score interpretations and the same tool inventory twice; not 2: there is no padded explanation of concepts Claude already knows, and each section does carry substance.

Actionability

There is real concrete guidance — exact tool names, explicit thresholds (ClinGen HI score 3, pLI >= 0.9, >=70% reciprocal overlap, >=1% BA1) and a report file-naming pattern — but the executable specifics are systematically deferred to 'ANALYSIS_PROCEDURES.md' ('implementation pseudocode'), 'CLASSIFICATION_GUIDE.md' ('scoring tables, ACMG code details'), and 'REPORT_TEMPLATE.md', none of which exist in the bundle, so the promised pseudocode, scoring breakdowns, and template are unavailable. Not 4: the missing key details go beyond minor gaps — no example tool invocation, no worked scoring computation, and the deferred files are dangling; not 2: it gives far more than high-level hints, with specific tools and quantitative cutoffs for each phase.

Workflow Clarity

The seven-phase workflow is clearly sequenced with a per-phase goal, an upfront reasoning checklist ('Work through these questions in order'), and specific checkpoints like 'PM2 requires absence from population databases at >=70% reciprocal overlap' and 'PS2 requires confirmed de novo status (check parental genotypes if available)'. Not 5: there are no explicit validate-then-proceed or error-recovery loops (e.g., what to do when ClinGen returns no data, or how to reconcile a high quantitative score with an ACMG benign call); not 3: the sequence is coherent, ordered, and includes several concrete gating conditions rather than only implicit checkpoints.

Progressive Disclosure

The in-body structure is good — phases stay lean and point outward with clearly signaled, one-level-deep references ('For SV type definitions, scoring tables, and ACMG code details, see CLASSIFICATION_GUIDE.md' plus a 'Reference Files' index) — but the bundle contains no files at all: CLASSIFICATION_GUIDE.md, ANALYSIS_PROCEDURES.md, REPORT_TEMPLATE.md, and EXAMPLES.md are all absent, so every reference is dangling and navigation dead-ends. Not 4: 'references mostly clear' requires the referenced files to actually exist — here the split is designed but not delivered; not 2: nothing is deeply nested or buried, and no bulk content that belongs in separate files is inlined — the problem is missing files, not monolithic structure.

Specificity

The description lists multiple concrete actions with full domain coverage: 'Applies ACMG-adapted criteria with ClinGen HI/TS dosage scores, gnomAD frequencies, and ClinVar evidence', 'Produces 5-tier classification with explicit per-criterion evidence', and 'CNV pathogenicity calls' / 'breakpoint analysis'. Not below 4: coverage spans interpretation inputs, method, and output; not applicable above since 5 is the top anchor and this matches it — every action is concrete, none generic.

Completeness

Both halves are explicit: 'what' ('Applies ACMG-adapted criteria... Produces 5-tier classification with explicit per-criterion evidence') and 'when' via the literal trigger clause 'Use for clinical genomics SV review, dosage-sensitivity assessment, breakpoint analysis, and CNV pathogenicity calls'. Not 4: the 'when' is explicit and enumerates concrete trigger phrases rather than being only generally stated; the explicit 'Use for' clause satisfies the cap rule without issue.

Trigger Term Quality

Good keyword coverage including the natural synonyms users would say: 'structural variant (SV)', 'deletions, duplications, inversions, translocations', 'CNV', 'clinical genomics SV review', 'dosage-sensitivity assessment'. Not 5: a few natural phrasings users actually use are missing (e.g. 'copy number variant', 'chromosomal rearrangement', a concrete question form like 'is this deletion pathogenic'); not 3: coverage already spans multiple synonym families (SV/CNV/structural variant) rather than just a few relevant keywords.

Distinctiveness Conflict Risk

It carves a clear niche — germline structural-variant clinical interpretation — with distinct triggers (SV, CNV, ClinGen dosage scores, breakpoint analysis) and minimal overlap risk; the closely related SNV/small-indel skill is explicitly a different territory. Not 4: even against a sibling variant-interpretation skill, the SV/CNV/dosage-sensitivity framing keeps conflict risk minimal rather than merely minor.