Local, file-based MCP exchange for DayVault; no model calls or direct database writes.
Three tools with complete JSON schemas and tool annotations present. Descriptions are in Chinese and moderately detailed (50-150 chars), meeting minimum standards. All tools have proper input schemas with type definitions and constraints. However, parameter descriptions are absent, the schema defines structure but LLMs cannot read JSON Schema directly and must infer parameter meaning from names alone. Tool naming follows verb_noun pattern (get_, list_, propose_) which is good. Error handling is present but generic. No output schema documentation visible.
读取用户选定目标的导出上下文。不是实时数据,不含其他目标或私人备注。
查看此目标导出的有效事项;可按状态筛选。不会读取 App 数据库。
在固定本地目录生成计划草案文件,仅限新建日期型事项。用户须在 DayVault 内预览确认;不会直接写入日程。
Parameter descriptions missing. Schema defines 'status' enum but LLMs cannot parse JSON Schema, they need explicit text descriptions explaining what each enum value means and when to use it.
Output schemas not documented. Tools return JSON but LLMs have no specification of response structure, field types, or what data to extract for downstream use.
Descriptions are in Chinese only. While culturally appropriate, descriptions lack English translations, limiting accessibility for non-Chinese-speaking developers and LLMs trained primarily on English.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2025-06-18+ | v2 |
Error messages are generic. safeError() function returns untranslated error text without recovery guidance. LLMs cannot determine if errors are retryable, user-fixable, or fatal.
Tool annotations present but incomplete. readOnlyHint and destructiveHint are set, but idempotentHint and openWorldHint values suggest these are static declarations rather than runtime-validated properties.