Content
92%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-engineered operational skill: every decision point maps to a concrete CLI command, destructive operations are gated by explicit user confirmation, and per-error retry semantics are specified. Progressive disclosure is exemplary, with a real one-level reference bundle; only minor wording repetition around identity selection keeps it from full marks on conciseness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and directive throughout — no conceptual padding or explanations of what Feishu/wiki is. Minor redundancy: the '--as user vs --as bot' identity caveat is repeated across several 快速决策 bullets and could be consolidated into the 身份选择 section. | 4 / 5 |
Actionability | Fully executable, copy-paste-ready commands with flags and JSON payloads: 'lark-cli wiki +node-get --node-token \'<wiki_url>\' --as user --format json', 'lark-cli contact +search-user --query "<姓名/邮箱/手机号>" --format json', and the raw department-search API call with explicit --params/--data JSON. Common cases (member resolution by user/chat/app/department, node listing, title rename) are all covered with concrete commands. | 5 / 5 |
Workflow Clarity | Multi-step flows with explicit validation checkpoints: delete-space requires resolving the real space_id, listing candidates, and '由用户明确选定一个 space_id 再执行' before the --yes delete; error handling is differentiated (invalid_parameters/not_found/permission_denied → fix per hint, do not retry; only rate_limit → backoff). Destructive operations have confirmation gates, so no cap applies. | 5 / 5 |
Progressive Disclosure | SKILL.md is a routing overview with a shortcut table linking to 13 references/*.md files (all verified present, one level deep), each keyed to a single operation. Detailed per-command content lives in those files; the body keeps only decisions, routing rules, and the API surface summary. Cross-skill pointers (../lark-shared, ../lark-drive) are clearly signaled and appropriate for a multi-skill bundle. | 5 / 5 |
Total | 19 / 20 Passed |