Content
86%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.
An exemplary lean skill body: a tight five-step workflow with concrete commands, a compatibility checklist, and proper deferral of detail to a real, well-structured reference file. The only meaningful gap is the absence of failure-handling guidance after the test run.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Thirteen lean lines with zero padding: no concepts Claude already knows are explained, every sentence instructs ("阅读对应 README、POM、相似接口、实现与测试", "执行 mvn -pl <module> -am test"). Every token earns its place. | 5 / 5 |
Actionability | Concrete, executable guidance throughout: real commands ("mvn -pl <module> -am test", "git diff --check"), concrete reading targets (README、POM、相似接口), a named PR target branch ("develop"), and parameterized placeholders. Not a 5 because some specifics (where to find Service/Bean templates, exact test-module locations) are deferred without inline pointers. | 4 / 5 |
Workflow Clarity | A clearly sequenced numbered 1–5 workflow ending in explicit verification commands (tests + "git diff --check"). Not a 5 because there is no error-recovery feedback loop (e.g. what to do when tests fail), and not below 4 since checkpoints are explicit rather than implicit. The destructive/batch cap does not apply — this is code contribution, not a batch/destructive operation. | 4 / 5 |
Progressive Disclosure | The body is a compact overview that delegates detailed conventions to a clearly signaled, one-level-deep reference ([贡献约定](references/contribution.md)), which exists on disk and is itself well organized with no further nesting. This matches the anchor-5 structure exactly. | 5 / 5 |
Total | 18 / 20 Passed |