Content
63%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 well-built router skill: clear navigation structure, decisive recommendations, and actionable pointers with no concept padding. Its weaknesses are systematic duplication across four routing sections and references that do not resolve within the provided bundle.
Suggestions
Consolidate the redundant routing sections — Example Invocations largely restates the Decision Tree, and Critical Patterns repeats the Quick Reference; keep one table and one decision path.
Ensure the referenced targets (skills/swift-testing.md, skills/ui-testing.md, etc.) ship with the bundle, or state explicitly that they are sibling-suite paths, so the router's one-level-deep references resolve.
Drop the dot-graph Decision Tree in favor of the numbered list alone — they encode the same routing and the graph adds tokens without new information.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The same routing facts repeat across four sections: "skills/swift-testing.md" is pointed to in the Quick Reference, Decision Tree, Critical Patterns, and Example Invocations ("→ Read: `skills/swift-testing.md`"), and Example Invocations largely restates the Decision Tree. Mostly efficient with no concept over-explanation, but roughly a quarter of the body is duplicated routing that could be tightened. | 3 / 5 |
Actionability | Concrete, executable navigation throughout: exact file paths ("skills/swift-testing.md"), agent names, a runnable command ("Launch `testing-auditor` agent or `/axiom:audit testing`"), and a decisive comparison table ("Unit tests (logic, models) | Swift Testing"). As an instruction-only router it has no code, which is not penalized, but the guidance itself is pointers rather than copy-paste content. | 4 / 5 |
Workflow Clarity | Routing is clearly sequenced via the numbered decision list and the symptom-to-reference table, with conditional checkpoints in the examples ("Check codebase: XCUIApplication/XCUIElement? → skills/ui-testing.md; Check codebase: @Test/#expect? → test-failure-analyzer"). Not a 5 because the dot graph duplicates the numbered list and agent-availability fallbacks live only in a dense blockquote. | 4 / 5 |
Progressive Disclosure | Structure is exemplary — one-level-deep references, clearly signaled, with a dedicated navigation table. However, scored against the actual bundle: the referenced targets (skills/swift-testing.md, skills/testing-async.md, skills/ui-testing.md, skills/xctest-automation.md, skills/ui-recording.md) do not exist in this bundle (no references/ or sibling skills/ directory), so every routing target is unverifiable. | 3 / 5 |
Total | 14 / 20 Passed |