Lean and information-dense with no concept padding — compact tables for parameter corrections, evidence tiers, and fallback chains; every token earns its place and Claude's competence is assumed.

Fully concrete guidance with exact tool calls and parameters (e.g. `Orphanet_search_diseases(operation="search_diseases", query=keyword)`, `FAVOR_annotate_variant("chr-pos-ref-alt")`) plus a WRONG/CORRECT parameter table and numeric scoring rules.

Clear 8-phase sequence (Phase 0–7) with key phase details, explicit fallback chains for error recovery, and evidence-grading tiers acting as validation checkpoints.

References are one-level-deep and clearly signaled with descriptions, but the actual bundle structure breaks this: DIAGNOSTIC_WORKFLOW.md, REPORT_TEMPLATE.md, and CHECKLIST.md are missing (only scripts/clinical_patterns.py exists), so detail that should be offloaded remains inline.

Lists multiple concrete actions — "HPO term matching to candidate diseases (Orphanet, OMIM), gene panel prioritization, ACMG variant interpretation, and structure-based variant analysis" — matching the anchor for multiple specific concrete actions.

Clearly answers both what (the four concrete diagnostic actions) and when, with an explicit "Use for..." trigger clause listing three triggering scenarios.

Natural trigger terms a target user would say appear in the use clause — "diagnostic odyssey assistance, phenotype-to-disease ranking, and genetic-counseling differential generation" — alongside domain terms like "rare disease" and "differential diagnosis".

Occupies a clear niche (rare disease differential diagnosis) with distinct triggers unlikely to fire for unrelated skills.