Content
61%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 solid, project-specific reference: commands and Jest configs are accurate and copy-paste ready, and the connector/handler/nock examples encode real conventions (e.g., matchedContactInfo, sqlite::memory: setup). The main costs are generic Jest boilerplate (unit-test skeleton, best-practices list) that adds tokens without new information, and the absence of any sequenced workflow or validation guidance for the test-write/run/coverage loop.
Suggestions
Cut the generic 'Unit Test Example' skeleton and 'Best Practices' list (basic Jest knowledge Claude already has), keeping only project-specific conventions like the matchedContactInfo return shape and the +1 'create new' option quirk.
Add a short sequenced workflow with a checkpoint, e.g.: write test → run `npm run test --workspace=@app-connect/core` → confirm nock assertions (`nock.isDone()`) and `jest.resetAllMocks()` before moving on.
Consider moving connector-specific test patterns (Pipedrive example) into a references/ file to slim the main SKILL.md to run/config/mocking essentials.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Most sections are lean code blocks with minimal prose, but the "Unit Test Example" (describe/it/beforeEach skeleton with Arrange-Act-Assert) and the "Best Practices" list ("Isolate tests", "Clean up", "Descriptive names") re-teach Jest fundamentals Claude already knows. This matches the 3 anchor (mostly efficient with some unnecessary explanation) better than the 4 anchor, since two whole sections could be cut. | 3 / 5 |
Actionability | Concrete, runnable commands ("npm run test --workspace=@app-connect/core", "npx jest --watch") and project-specific examples for nock mocking, handler testing with supertest, and the Pipedrive connector are executable. The generic unit-test example uses placeholders ("functionUnderTest", "const input = { ... }") rather than complete code, which is the minor gap separating it from the 5 anchor; it is far above the 3 anchor's pseudocode level. | 4 / 5 |
Workflow Clarity | The body is a pattern catalog organized by topic rather than a sequenced workflow: there is no "write test → run → check coverage" guidance or validation checkpoints (e.g., verify mocks cleaned / nock assertions before committing). It sits at the 3 anchor (content present, checkpoints missing) rather than 2 because each section's usage is individually unambiguous. | 3 / 5 |
Progressive Disclosure | The body (~270 lines) is a single well-sectioned file with no bundle files, clear headers (Running Tests, Test Patterns, Test Utilities, Coverage), and every path referenced (tests/setup.js, tests/fixtures/callLog.js, packages/core/jest.config.js) matching the documented project layout. It does not reach 5 because content like the connector/handler test patterns is sizable enough that a reference split could reduce the main file further, but the inline placement is cohesive and easy to navigate. | 4 / 5 |
Total | 14 / 20 Passed |