Content
71%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 well-structured as an overview with an excellent single one-level-deep reference split and a workflow that has genuine validation gates (compile-before, confirm-before-code, verify-after). Its main weakness is redundancy: preconditions, compile checks, and 'read the reference' are each restated two or three times across the Constraints, Precondition, and Workflow sections, and the trigger list duplicates the frontmatter description.
Suggestions
State each invariant once: merge the 'Preconditions' paragraph into the 'PRECONDITION' constraint and drop the repeated compile/stop instructions from workflow step 1, keeping the canonical command only in the Constraints checklist.
Remove the 'When to use this skill' section — it repeats the five trigger phrases verbatim from the frontmatter description, which is already loaded into context whenever the skill is considered.
Reduce the three separate 'read the reference' mentions (BEFORE APPLYING, step 2, Reference section) to a single well-signaled pointer in the workflow step where the reference is actually consumed.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body repeats itself across sections: the precondition about maintainer-sanitized scenario facts appears in 'Preconditions', the 'PRECONDITION' constraint, and workflow step 1; 'run ./mvnw compile' appears in MANDATORY, SAFETY, and step 1; 'read the reference' is stated in Constraints ('BEFORE APPLYING'), step 2, and the Reference section; and 'When to use this skill' re-lists the five trigger phrases verbatim even though the frontmatter description already carries them. It fits the anchor 'mostly efficient but includes some unnecessary explanation or could be tightened' — no concept over-explanation, but the structural redundancy keeps it below score 4, which requires only minor trims. | 3 / 5 |
Actionability | Concrete, executable guidance is present throughout: exact commands ('./mvnw compile or mvn compile', './mvnw clean verify'), specific Maven coordinates ('rest-assured and wiremock-standalone', 'maven-failsafe-plugin'), naming conventions ('*AT' classes, '@acceptance' tags), and concrete mechanisms ('@BeforeAll coordinate propagation via System.setProperty'). It sits at 'mostly executable guidance; concrete code or commands with minor gaps' — the Java/test code itself is deliberately deferred to the reference, and small gaps remain (e.g., no failsafe plugin configuration snippet inline), so it does not reach the copy-paste-ready score-5 anchor. | 4 / 5 |
Workflow Clarity | The four-step workflow is clearly sequenced with real validation checkpoints: compile gate before changes, confirm-scenarios-before-coding gate, and a full 'clean verify' after, each with an explicit stop condition ('If compilation fails, stop immediately'). This satisfies 'clear sequence with most checkpoints present'. It falls short of the score-5 anchor, which requires feedback loops for error recovery — the body mandates stopping on failure rather than a fix-and-retry cycle — so score 5 does not fit better than 4. | 4 / 5 |
Progressive Disclosure | The SKILL.md body is a lean ~60-line overview while the 527-line detailed reference (references/133-java-testing-acceptance-tests.md) exists, is real, and is exactly one level deep with no further nested references; the body signals it clearly in workflow step 2 and a dedicated '## Reference' section with a markdown link, plus a 'What is covered' index. This matches the anchor 'clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation'. | 5 / 5 |
Total | 16 / 20 Passed |