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 highly actionable, well-structured skill body: every workflow is executable end-to-end with a shared curl helper, and API detail is properly offloaded to a clearly referenced bundle file. The main improvement areas are trimming the multi-platform UTF-8 section and making response validation explicit in the workflows.
Suggestions
Move the per-platform UTF-8 conversion snippets (Python/Node/Unix/PowerShell) into references/api.md and keep only a one-line rule in SKILL.md, e.g. "写入前确保 UTF-8，转码方法见 references/api.md".
Add an explicit validation step to each workflow, such as checking the response error code is 0 before extracting docid/folder_id, and what to do on 100002/20004 (credential issues).
State the privacy rule (只展示标题和摘要，禁止展示正文) once — it currently appears in both the API 调用模板 section and 注意事项.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The decision table, curl workflows, enum and error tables are dense and operational, but the ~50-line UTF-8 section covering four platforms and the privacy rule stated in both "API 调用模板" and "注意事项" are trimmable over-explanation. It is above the 3 anchor because nearly every token is non-obvious domain specifics rather than concepts Claude already knows. | 4 / 5 |
Actionability | The ima_api helper function, credential precheck script, and copy-paste-ready JSON bodies for every workflow (search, browse, create, append) are fully executable and cover the common cases, matching the top anchor. | 5 / 5 |
Workflow Clarity | Workflows are clearly sequenced (search → get_doc_content; list folders → list notes) with a credential precheck and an error-code table giving recovery actions like splitting large writes into multiple append_doc calls. It falls short of 5 because per-step validation (e.g., checking the response error code before parsing docid) is implicit rather than an explicit checkpoint. | 4 / 5 |
Progressive Disclosure | The body is well sectioned and points to references/api.md (a real, one-level-deep file) twice with clear signaling, with bulk API detail appropriately split out. It sits below the 5 anchor because the 核心响应字段 summary and the multi-platform UTF-8 conversion guide are inlined where an ideal overview would keep the body leaner. | 4 / 5 |
Total | 17 / 20 Passed |