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

Annotations already establish readOnly/closed-world/non-destructive, so the safety profile is covered. The description adds the substantive behavioral constraint that the two AUCs must come from the same subjects (paired design) and that the output is the test of a difference, which is the key thing an agent must get right before calling it. It omits assumptions such as binary labels being 0/1, minimum sample requirements, and error behavior, so it is not exhaustive.

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 labeled structure (USE WHEN / ALIASES / RETURNS) is front-loaded and easy to scan. However, the opening line duplicates the tool title verbatim in both English and Chinese, and the six-item alias list is keyword padding rather than information an agent needs to invoke the tool correctly.

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?

With no output schema, the description carries the burden of explaining returns, and 'RETURNS: 两 AUC 及差值 P 值' covers only the bare minimum (two AUCs plus a difference p-value) — no direction of the difference, confidence interval, or per-AUC standard errors. For a simple three-array statistical tool this is adequate but thin, especially given the mixed-language phrasing that may slow parsing.

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?

Schema coverage is 100%, so per-dimension field meaning is already documented (labels 0/1, scores for metric 1 and metric 2). The description goes beyond the schema by implying the three arrays must be subject-aligned and equal length, a constraint the schema never states. It still does not spell out that constraint explicitly or discuss score scale/ties.

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?

States a specific named statistical test (DeLong) with its exact operation: comparing two ROC AUCs from the same subjects to test whether the difference is significant. This cleanly separates it from the sibling roc_auc (single-AUC estimation) and from generic metric tools, so an agent can identify it without opening the schema.

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?

The 'USE WHEN' clause gives an explicit trigger condition: two AUCs measured on the same subjects, testing for a significant difference. It does not state when NOT to use it (e.g., independent/unpaired groups, where a different test would be needed) or name an alternative tool, so it falls short of a full routing rule.

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