Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured skill body: a command-selection table, executable worked examples, focused caveats, and a clear out-of-scope section. Content is split appropriately across real reference files with clean navigation.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean body that assumes Claude's competence — a decision table, short concrete examples, and bullet caveats with no preamble explaining what Feishu/Lark is, matching the lean-and-efficient anchor. | 3 / 3 |
Actionability | Provides fully executable, copy-paste-ready commands with real flags (e.g. 'lark-cli contact +search-user --query "张三" --has-chatted --as user' and the batch_query invocation with explicit --params/--data JSON), not pseudocode. | 3 / 3 |
Workflow Clarity | The search→send flow is sequenced ('先搜,确认 open_id,再发') with an explicit disambiguation checkpoint ('搜索命中多条…把候选列给用户挑;不要擅自选第一条'); the batch operation is a non-destructive read so the missing feedback loop does not cap it. | 3 / 3 |
Progressive Disclosure | Overview body points to two real, one-level-deep reference files (lark-contact-search-user.md, lark-contact-get-user.md) via clearly signaled links and defers sibling domains to other skills, with no nested reference chains. | 3 / 3 |
Total | 12 / 12 Passed |