Content
78%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-organized, highly actionable body with an excellent executable example and useful API contracts. Its two real weaknesses are the orphaned references/api_reference.md bundle file (never linked, so its advanced examples and troubleshooting are undiscoverable) and mild capability-list redundancy across three sections.
Suggestions
Add a clearly signaled, one-level-deep reference to the bundle, e.g. an "## Advanced" or "## API reference" section with "See [api_reference.md](references/api_reference.md) for detailed method signatures, advanced examples, and troubleshooting" — this also surfaces the troubleshooting guidance the body lacks.
Trim redundancy by merging the "When to Use" bullets (which restate the frontmatter description) into a single short triggers section, and cutting the repeated PCoA/PERMANOVA/ANOSIM/Mantel list that appears in When to Use, Key Features, and Implementation Details.
Remove the duplicate dependency declaration — either the Dependencies section or the "pip install scikit-bio numpy pandas" comment in the code block, not both.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient overall: it assumes domain competence (no explanations of what DNA, diversity, or ordination are) and every section carries API-contract information Claude would not know, e.g. "Count inputs should be non-negative integers representing abundances (not relative frequencies)". Minor trimmable redundancy keeps it at anchor 4 rather than 5: "When to Use" bullets, "Key Features" bullets, and Implementation Details partially restate the same capability list (e.g., PCoA/PERMANOVA/ANOSIM/Mantel appear in three places), and the dependency list plus "pip install" comment duplicate each other. | 4 / 5 |
Actionability | The Example Usage block is fully executable and copy-paste ready — complete imports (skbio, DNA, TreeNode, alpha_diversity, beta_diversity, pcoa, permanova), inline sample data, and real API calls with correct signatures (alpha_diversity("shannon", counts, ids=sample_ids), beta_diversity("braycurtis", ...), pcoa(dm), permanova(dm, grouping=grouping, permutations=99), TreeNode.read([newick]), tree.shear(...)). It covers the common cases (sequence ops, alpha/beta diversity, ordination, PERMANOVA, tree I/O), matching anchor 5. Implementation Details adds the concrete return-type contracts that make follow-up work reliable. | 5 / 5 |
Workflow Clarity | The example is explicitly sequenced (numbered sections 1-5: sequence manipulation → counts/diversity → ordination → PERMANOVA → tree I/O) and Implementation Details explains the data flow (counts → DistanceMatrix → ordination/tests), giving a clear progression with minor gaps — matching anchor 4 ("Clear sequence with most checkpoints present; minor validation gaps"). No validation checkpoint is required because nothing here is destructive or batch-risky, so the cap of 3 does not apply; a 5 would need explicit error-recovery guidance (e.g., what to do when phylogenetic metrics fail from mismatched tree/feature IDs), which is only in the unreferenced troubleshooting section of the bundle. | 4 / 5 |
Progressive Disclosure | The bundle contains references/api_reference.md (748 lines of detailed API docs, advanced examples, and a Troubleshooting section), but the SKILL.md body never mentions or links it — the only link in the body is an external GitHub source URL. Per the guideline to score against the actual bundle structure, the reference is present yet completely unsignaled, and API-contract detail that could live in the reference is inlined in Implementation Details. This matches anchor 3 ("references present but not clearly signaled; content that should be separate is inline"); it is above anchor 2 because the body itself has good section structure and is not a monolithic wall. | 3 / 5 |
Total | 16 / 20 Passed |