Content
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill provides excellent actionable guidance with clear workflows, executable code examples, and explicit verification checkpoints. However, it is severely bloated with repetitive motivational content—the same arguments against skipping TDD are made in at least 4 different sections (Why Order Matters, Common Rationalizations table, Red Flags list, and TDD Prevents These Anti-Patterns). Claude doesn't need persuasion; it needs instructions. The anti-patterns section, while valuable, should be a separate file to improve progressive disclosure.
Suggestions
Remove the 'Why Order Matters' section and 'Common Rationalizations' table entirely—they repeat the same points and Claude doesn't need motivational arguments, just rules to follow.
Extract the 'Testing Anti-Patterns' section into a separate ANTI_PATTERNS.md file and reference it from the main SKILL.md to improve progressive disclosure and reduce document length.
Consolidate the 'Red Flags' list—many items are restatements of the same principle ('Code before test,' 'Test after implementation,' 'Tests added later,' 'I already manually tested it' all mean the same thing).
Remove the dot graph definition—Claude cannot render Graphviz diagrams, and the Red-Green-Refactor cycle is already clearly explained in the subsequent sections.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose and repetitive. The same points about 'delete code and start over,' 'tests after prove nothing,' and 'sunk cost fallacy' are restated across multiple sections (Why Order Matters, Common Rationalizations, Red Flags). The anti-patterns section repeats TDD principles already covered. The motivational/philosophical content (explaining why TDD matters, countering rationalizations) is extensive and largely unnecessary for Claude, who doesn't need persuasion—just instructions. | 1 / 3 |
Actionability | Provides fully executable TypeScript code examples for both good and bad patterns, concrete bash commands for verification steps, and specific examples for bug fixes. The code is copy-paste ready and demonstrates real behavior testing patterns clearly. | 3 / 3 |
Workflow Clarity | The Red-Green-Refactor cycle is clearly sequenced with explicit verification checkpoints at each stage (Verify RED, Verify GREEN). Includes feedback loops (test errors → fix → re-run, test passes unexpectedly → fix test). The verification checklist at the end provides a comprehensive gate before marking work complete. The bug fix example walks through the complete cycle. | 3 / 3 |
Progressive Disclosure | The content is a monolithic document combining two substantial topics (TDD methodology and Testing Anti-Patterns) in a single file. The anti-patterns section could be a separate referenced file. No bundle files exist to offload detailed content. The document is well-structured with clear headers but is overly long for a single SKILL.md. | 2 / 3 |
Total | 9 / 12 Passed |