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.
The body delivers genuinely useful, opinionated guidance (decision rule, naming conventions, testing seams) with good section structure, but it spends tokens restating standard hexagonal-architecture knowledge and lacks an explicit design sequence or verification step. The unreferenced references/disclaimer.md file is a dangling bundle artifact.
Suggestions
Add a validation checkpoint to the workflow, e.g. a final 'verify dependencies point inward: no adapter/framework types imported by domain or application code' step, so the design procedure has a concrete pass/fail check.
Trim restatements of standard hexagonal knowledge (layer contents like 'Entities, Value Objects, Domain Services', 'Dependencies flow INWARD only', the anemic-domain definition) and keep only the project-specific rules and conventions.
Either reference references/disclaimer.md from the body (e.g., a status note near the top) or remove it from the bundle so every bundle file is discoverable from SKILL.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient with good use of tables/lists, but it restates concepts Claude already knows: "Contains Entities, Value Objects, Domain Services", "Dependencies flow INWARD only", the anemic-domain description, and general hexagonal layering — more than minor over-explanation, matching the 'mostly efficient but includes some unnecessary explanation' anchor. | 3 / 5 |
Actionability | Concrete, applicable guidance dominates: the decision rule ("Does it do I/O or run out-of-process?"), naming patterns with examples ("MemberView.from(member)", "request.asMember()", "MemberDbo"), and a layer-to-test-type mapping. Minor gaps remain — several anti-pattern fixes are high-level with no example, e.g. "Use wrapper objects that can evolve without breaking signatures". | 4 / 5 |
Workflow Clarity | The core decision rule implies a procedure but there is no explicit step sequence and no validation checkpoint for verifying the boundary holds (e.g., checking that no adapter types leak into the domain) — the sequence is implicit and checkpoints are missing, matching anchor 3. | 3 / 5 |
Progressive Disclosure | Sections are well-organized and the ~82-line body is reasonably self-contained, but the bundle contains references/disclaimer.md which is never referenced or linked from the body — a dangling file that represents a minor organization gap against the actual bundle structure. | 4 / 5 |
Total | 14 / 20 Passed |