Content
60%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 an effective, well-signaled router with concrete per-task file routing and explicit conflict-resolution rules between sibling skills. Its main weakness is redundancy — the routing table is restated four times in different formats — plus a stub subagent-delegation section and references that cannot be verified against the bundle.
Suggestions
Collapse the dot-graph decision tree, the numbered decision list, and the Example Invocations into the single Quick Reference table — the routing mapping currently appears four times and could roughly halve the file.
Cut or populate the "Cursor Subagent Routing" section: it instructs delegation to "the appropriate Cursor subagent" but names none except liquid-glass-auditor, which already appears in the decision tree.
Trim "Critical Patterns" to one-line summaries or fold it into the Quick Reference, since it restates the scope of each reference file the table already routes to.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The same task-to-reference routing is repeated four times — Quick Reference table, dot-graph decision tree, numbered decision list, and Example Invocations — and "Critical Patterns" re-summarizes the same files again, so the ~185-line body carries several redundant/padded sections. It avoids explaining concepts Claude already knows (better than anchor 1), but the duplication goes beyond anchor 3's "could be tightened". | 2 / 5 |
Actionability | Routing is concrete and executable for an instruction-only skill: each symptom maps to a specific file ("See skills/hig.md", "skills/sf-symbols-ref.md (Part 11)", "/skill axiom-swiftui"), and Example Invocations map real user phrasings to files. Gaps: the "Cursor Subagent Routing" section delegates to unnamed subagents (only liquid-glass-auditor is ever named), keeping it below anchor 5. | 4 / 5 |
Workflow Clarity | The numbered decision tree with sub-steps (1a, 2b, 3a) plus the Conflict Resolution section gives a clear, mostly unambiguous routing sequence with explicit precedence rules between overlapping skills. No destructive or batch operations apply, so no validation cap; the underspecified subagent-delegation path is the minor gap versus anchor 5. | 4 / 5 |
Progressive Disclosure | The body is a proper overview pointing to one-level-deep, clearly signaled references ("See skills/liquid-glass.md"), which is the right structure. However, none of the referenced skills/*.md files exist in the provided bundle, and reference styles are inconsistent (file paths, "/skill" invocations, named suites like "axiom-payments"), so navigation is not fully verifiable — anchor 4's "minor organization gaps" rather than anchor 5's "easy navigation". | 4 / 5 |
Total | 14 / 20 Passed |