Content
82%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-constructed skill body: executable WebFetch and CLI commands, a disciplined five-step discovery workflow with safety confirmation for write/delete operations, and clean delegation of shared concerns to lark-shared. The weaknesses are minor — a duplicated translation note, a hedged dry-run suggestion, and no explicit error-handling feedback loop.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean — a doc-library tree diagram, a compact entry-URL table, five terse steps, and two worked examples — with no padding explaining concepts Claude already knows. Minor trimmable redundancy keeps it off the 5 anchor: the English-translation instruction appears twice ('如果用户使用英文交流,需将文档内容翻译为英文后输出' and again '如果用户使用英文交流,将以上所有内容翻译为英文'), and the brand URL rules are stated twice (table + Step 2 bullets). | 4 / 5 |
Actionability | Fully executable throughout: each WebFetch step includes a concrete extraction question, the `lark-cli api` section gives copy-paste syntax for GET/POST/PUT/DELETE, and the two scenarios are complete end-to-end command sequences ('lark-cli api POST /open-apis/im/v1/chats/oc_xxx/members --data … --params …'). This matches the 5 anchor — copy-paste-ready commands with specific examples covering the common cases. | 5 / 5 |
Workflow Clarity | Five strictly sequenced steps with an explicit anti-skip directive ('严格按以下步骤逐层检索,不要跳步或猜测 API'), a capability checkpoint at Step 1, and destructive-op confirmation in the safety rules ('写入/删除类 API 调用前必须确认用户意图') — so the destructive-operation cap at 3 does not apply since validation is present. It falls short of the 5 anchor because the dry-run safety valve is hedged ('如支持') and there is no error-recovery loop (nothing on handling error codes or failed calls), matching 'clear sequence with most checkpoints; minor validation gaps'. | 4 / 5 |
Progressive Disclosure | Well-organized sections with clearly signaled one-level-deep cross-skill references — auth and safety detail delegated to [lark-shared](../lark-shared/SKILL.md) rather than duplicated, and a 参考 section listing both links with purpose. No bundle files exist to split content into; the inline material (tree diagram, two scenarios) is appropriately sized for SKILL.md. The 5 anchor is not fully met only because the cross-skill paths sit outside this bundle and cannot be verified, and the scenario examples could arguably live in a reference file — minor organization gaps per the 4 anchor. | 4 / 5 |
Total | 17 / 20 Passed |