Content
81%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.
The body is highly actionable with a clear, validated workflow and exemplary reference routing, but it carries redundant sections and references a tests/data/ bundle that does not actually exist. Tightening the duplicated validation content and either shipping the test data or removing those references would lift the skill's overall quality.
Suggestions
Either bundle the referenced tests/data/expression_matrix_merged.csv and tests/data/sample_info.csv or remove the tests/data/ row from the file table, the Testing example commands, and the 'Test data provided' checklist item — the paths currently point at files that do not exist.
Merge the 'Input Validation' section with 'Input Format', since both state the same acceptance criteria (2 groups, 2 batches, 2 samples per group/batch, unique non-empty IDs).
Remove the 'Implementation Checklist' section (author-facing, not agent-facing) and the 'Last updated / Version' footer to reduce token spend without losing operational value.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Most of the body is dense, operational tables (arguments, outputs, errors, examples), but the 'Input Validation' section restates requirements already in 'Input Format', the 'Implementation Checklist' is author-facing padding, and the prerequisites note duplicates the install command. This fits anchor 3 ('mostly efficient but some unnecessary explanation or could be tightened') better than anchor 4, since two whole sections are redundant rather than minor phrasing. | 3 / 5 |
Actionability | Fully executable, copy-paste-ready commands: the full CLI invocation with all flags, an argument table with defaults, concrete CSV input examples, three usage examples covering common cases, error codes with solutions, and test commands. This matches anchor 5. | 5 / 5 |
Workflow Clarity | Clear sequence (prerequisites -> run -> verify outputs -> report per the Agent Response Contract) with explicit validation (Step 1 'Validate Input' plus Testing validation commands) and a feedback loop for error recovery ('IF error persists, READ references/troubleshooting.md', plus the dependency-first troubleshooting note). This matches anchor 5 rather than 4, which would require missing checkpoints. | 5 / 5 |
Progressive Disclosure | The 'When to Read External Files' situation-based table provides well-signaled, one-level-deep references that all exist (references/algorithm.md, scripts/main.R, references/troubleshooting.md, references/cli-guide.md). However, 'tests/data/' is referenced in the table and the Testing section (with specific filenames) and asserted as provided in the checklist, yet no tests/ directory exists in the bundle — a broken path that keeps this below anchor 5. | 4 / 5 |
Total | 17 / 20 Passed |