Content
81%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-structured, actionable workflow with strong validation and security gates for destructive operations. Tightening the credential-model prose and moving the large tool tables to a reference file would lift conciseness and progressive disclosure.
Suggestions
Consolidate the repeated secret-handling warnings into one canonical security section to reduce token redundancy.
Move the read/write/destructive tool tables into a references/TOOLS.md file and link to it from the body to improve progressive disclosure.
Add full argument signatures (e.g. send_message(to, cc, bcc, subject, body)) next to each write/destructive tool so examples are copy-paste ready.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and assumes Claude's competence (no basic email explanations), but the credential/security model and repeated secret-handling warnings add some length that could be tightened. | 4 / 5 |
Actionability | Concrete MCP tool names with parameters (create_mailbox(domain, username), check_auth, set_credentials) and exact registration URLs are given, though a couple of tool invocations omit full argument signatures. | 4 / 5 |
Workflow Clarity | A clearly numbered multi-step flow (第1步 / 第2a / 第2b / 第3步) with explicit validation checkpoints (check_auth 验证成功前不要调用其他邮件工具) and a user-confirmation gate before destructive delete_message/delete_thread. | 5 / 5 |
Progressive Disclosure | Well-organized into clear sections with a one-level-deep CONTACTS.md reference, but the inline credential model and tool tables could plausibly live in separate reference files given the skill's length. | 4 / 5 |
Total | 17 / 20 Passed |