Content
45%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 a router with clear routing logic and a defensible conceptual premise, but the skill is broken as bundled: both files the Router points to are missing, leaving only abstract framing with no actionable substance. Redundancy between the Shared Premise and Scope sections further dilutes an otherwise lean 30-line overview.
Suggestions
Include the referenced bundle files: every row of the Router table links to state-ownership.md or component-composition.md, and neither exists in the skill directory, so the skill currently delivers no substantive guidance when its links are followed.
Inline brief definitions of the core concepts the body depends on (Proximity Principle, Containment, Specialization) — or ensure the referenced files define them — so the overview is not built on undefined terms.
De-duplicate the Shared Premise and Scope sections: the co-location prohibition and the "current structure is not the design" warning each appear twice and could be stated once, tightening conciseness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The 30-line body is mostly lean, but the co-location prohibition appears twice ("Do not mechanically co-locate state and logic in the nearest individual consumer" in Shared Premise vs. "do not initially co-locate state in the nearest consumer" in Scope) and the "don't trust current structure" warning is likewise restated across sections. It is not 4 because this duplication exceeds minor over-explanation; not 2 because there is no padding that explains basics Claude already knows. | 3 / 5 |
Actionability | The body offers high-level decision framing ("Identify the UI unit and the component set that implements it first") and a routing table, but every concrete step and example is deferred to state-ownership.md and component-composition.md — and neither file exists in the bundle, so the specific steps to execute are missing entirely. Terms the guidance depends on ("Proximity Principle", "Containment and Specialization") are used but never defined in the body. It is not 3 because there is no partially-complete executable guidance present; the substantive content is wholly absent. | 2 / 5 |
Workflow Clarity | The Router gives an unambiguous decision procedure ("Route by the required design decisions, not terms in the user's request. Read only the necessary documents, and read both when the decisions overlap") with explicit ordering ("Read state-ownership.md first, then component-composition.md"), and Scope separates first-implementation from refactoring. It is not 5 because the identify-the-UI-unit step is scattered across sections rather than presented as a sequenced workflow; not 3 because for a routing skill the if/then logic and reading order are clear, and no validation checkpoints are needed for this non-destructive design-decision skill. | 4 / 5 |
Progressive Disclosure | Scored against the actual bundle structure: the router table is well-signaled and one level deep, but both referenced files (state-ownership.md and component-composition.md) do not exist anywhere in the skill directory — no references/, scripts/, or assets/ directories are present — so every reference is dangling and the content the skill discloses is absent. It is not 1 because the structure and signaling are clear rather than a monolithic or nested chain; not 3 because this is not merely imperfect organization — the entire referenced payload is missing. | 2 / 5 |
Total | 11 / 20 Passed |