Conciseness

The body is dense and efficient — tables of tool parameters, point-scoring rubrics, and parallelization groups with no explanations of concepts Claude already knows. However, the Critical Tool Parameters tables, Scoring Summary, and hardcoded "Known CIViC Gene IDs" list duplicate material the Reference Files table assigns to TOOLS_REFERENCE.md and SCORING_CRITERIA.md, so not every token earns its place. Fits the 4 anchor (efficient, minor instances that could be trimmed) rather than 5.

Actionability

Guidance is concrete and executable in spirit: exact tool names with required parameters ("action must be exactly 'search_studies'", "ALL 4 parameters... REQUIRED"), batch size of 10, a 0-100 scoring breakdown, and parallelization groups. It falls short of 5 because the body contains no worked search-query or code examples — those are deferred to EXAMPLES.md and TRIAL_SEARCH_PATTERNS.md, leaving minor gaps in copy-paste-ready guidance.

Workflow Clarity

The ten phases are clearly sequenced with inputs/outputs per phase, parallelization groups, and an Error Handling section (try/except per call, empty-result checks, fallback tools, documented failures) that serves as validation for the batch operations, so it does not hit the batch-operation cap of 3. It is not 5 because checkpoints are described in a separate section rather than integrated as explicit per-phase validation steps, and the "KEY PRINCIPLES" list is structurally broken — items 4-10 are orphaned under the "Molecular Matching Priority" heading, interrupting the sequence.

Progressive Disclosure

The body is an overview with a Reference Files table signaling seven one-level-deep references with their contents, which is good structure. It is not 5 because substantial reference material (the full tool parameter tables, CIViC gene ID lookup, detailed scoring tables) is inlined rather than split into those files, creating duplication, and none of the seven referenced files exist in the provided bundle to verify the links resolve.

Specificity

The description lists multiple concrete actions with comprehensive coverage: "Transforms a patient's molecular profile (mutations, biomarkers, expression) and clinical state into ranked clinical-trial recommendations with evidence tiers", "Searches ClinicalTrials.gov, the EU CTIS register... AND the ISRCTN registry", and "cross-references CIViC, OpenTargets, ChEMBL, and FDA labels" — all in consistent third-person voice. It clearly matches the 5 anchor (multiple specific concrete actions); there is no vagueness that would drop it to 4.

Completeness

It explicitly answers both questions: the "what" is stated concretely ("Transforms a patient's molecular profile... into ranked clinical-trial recommendations with evidence tiers") and the "when" is an explicit trigger clause ("Use for matching patients to trials by genotype, biomarker-driven trial selection, trial-eligibility scoring, and finding trials across the US, Europe, and the UK"). This matches the 5 anchor exactly; 4 would require a weaker or less explicit 'when' clause.

Trigger Term Quality

Natural phrases like "matching patients to trials", "biomarker-driven trial selection", "trial-eligibility scoring", and "finding trials across the US, Europe, and the UK" give good keyword coverage, but common user phrasings (e.g. "find clinical trials for my cancer", "trial search", "enrollment options") are missing and the term list leans technical (register and database names). This fits the 4 anchor (good coverage, a few natural terms missing) rather than 5 (comprehensive synonyms).

Distinctiveness Conflict Risk

It carves out a clear niche (AI-driven patient-to-trial matching for precision oncology and rare-disease care) with triggers specific to trial matching by genotype and biomarker, so overlap with adjacent bioinformatics skills (variant interpretation, drug safety, target validation) is minimal. Matches the 5 anchor (clear niche with distinct triggers); 4 would imply minor overlap risk that is not evident here.