Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

The description adds behavioral context beyond annotations: harvest is a long task taking seconds to minutes, semantic mode=deep is '更慢也更贵', and failure/empty results are not billed. Annotations declare destructiveHint=true, and while the description does not explain destructive behavior, it also does not contradict it; the provided latency and cost details justify the score.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The list format is organized and each sub-tool line is dense with information, but the description is quite long and contains a truncated sentence ('它回答的不是「搜到了什么」，而') that breaks the structure and indicates an editorial flaw. It is not as concise as it could be, and the incomplete sentence hurts overall quality.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

The description covers all 17 sub-tools with enough context to invoke them, and it provides cost/delay warnings. However, the crosscheck description is incomplete, so its core value is partially unclear. Additionally, there is no explanation of return values or output structure, and the 'params' object format is only loosely illustrated, leaving some ambiguity for an agent.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema covers both parameters, with 'tool' as an enum and 'params' as a generic object. The description substantially enriches parameter meaning by specifying expected inputs per sub-tool: URL inputs for lens/similar/read, natural-language queries for crawl/semantic, kind=patents for scholar, and full=true for deep. This goes well beyond the schema's single generic example.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly identifies the tool as a search dispatcher and distinguishes 17 sub-tools with specific verbs and objects (e.g., 'crawl: 爬一个站点的多个页面并取回正文', 'lens: 反向图搜'). It differentiates internal modes well, but the crosscheck entry contains a truncated sentence ('它回答的不是「搜到了什么」，而') that leaves its distinguishing purpose incomplete, preventing a perfect score.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicit usage guidance is provided for several sub-tools: harvest '只在 crawl 也拿不到时用', crosscheck is recommended for asserting concrete facts, semantic is suggested when keyword search fails, and forum is described as an independent evidence source from official docs. This clearly helps an agent select the right sub-tool.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.