Content
75%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 lean, well-organized process reference with concrete rules and a clearly sequenced red→green loop. Its main defects are that both files it references are absent from the bundle (breaking the disclosure chain) and that the loop's verification checkpoint is left implicit.
Suggestions
Ship tests.md and mocking.md in the skill bundle (or remove the references from SKILL.md): the body links to both, but neither file exists alongside SKILL.md, so the promised examples and mocking guidelines are unreachable.
Add a minimal worked example of one full red→green cycle (a named seam, the failing test, then the minimal implementation) directly in the body so the loop is concrete even without the external example file.
Make the loop's verification checkpoint explicit — e.g., "run the test and confirm it fails for the expected reason before writing any implementation" — rather than leaving the red-phase gate implicit in "Red before green".
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean, prescriptive doctrine ("No test is written at an unconfirmed seam", "One seam, one test, one minimal implementation per cycle") with only minor over-explanation of standard testing wisdom Claude already knows ("A good test reads like a specification… it survives refactors because it doesn't care about internal structure"). Not 5 because a few such passages could be trimmed; not 3 because padding is minor rather than a pattern. | 4 / 5 |
Actionability | Concrete, executable directives are present: "Before writing any test, write down the seams under test and confirm them with the user", a copy-paste question ("What's the public interface, and which seams should we test?"), and a concrete tautology example ("expect(add(a, b)).toBe(a + b)"). Not 5 because no worked red→green cycle example exists in the body and the examples are delegated to tests.md, which is not in the bundle; not 3 because the guidance given is specific and directive, not pseudocode or vague hints. | 4 / 5 |
Workflow Clarity | The loop is clearly sequenced with the failing test as a gate: "Red before green. Write the failing test first, then only enough code to pass it", plus the rule that changes to existing code start a new cycle. Not 5 because the verification checkpoint is implicit — it never says to run the test and confirm it fails for the expected reason before writing implementation; not 3 because the sequence and its checkpoints are mostly explicit and this is not a destructive/batch operation. | 4 / 5 |
Progressive Disclosure | The body is short and well-sectioned with clearly signaled, one-level-deep references ("See [tests.md](tests.md) for examples and [mocking.md](mocking.md) for mocking guidelines"). Held below 5 — and straining the 4 anchor — because neither referenced file exists in the bundle (no references/, scripts/, or assets/ directories are present), so navigation to the delegated examples and mocking guidance is broken; not 3 because the in-file structure and reference signaling are good and nothing that belongs in separate files is inlined. | 4 / 5 |
Total | 16 / 20 Passed |