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

Annotations already declare readOnlyHint=true, destructiveHint=false and openWorldHint=false, so the safety profile is covered and the description only needs to add operational context. It adds that the full diagnostic set is returned, but says nothing about how the threshold parameter affects the binarization of predictions or what the output structure looks like — meaningful gaps for a metrics tool with a tunable cutoff.

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 pipe-delimited keyword-block format is dense, front-loaded, and easy to scan, with the primary use case stated before aliases. The alias list is somewhat long and partly duplicates the metric names already listed, but every segment is short and functional.

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?

For a simple three-parameter statistics tool with full schema coverage and a read-only annotation profile, the description covers what it computes and the metric set returned. The only shortfall is the absence of any return-format detail, which matters slightly more because there is no output schema.

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 description coverage is 100% (labels, predictions, threshold all documented in Chinese), so the schema carries parameter meaning and a baseline of 3 applies. The description does not explain the threshold's default, its role in converting probabilities to 0/1, or the expected alignment between labels and predictions arrays.

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 names the resource (classification metrics) and enumerates the exact metric set returned (accuracy, sensitivity, specificity, PPV, NPV, F1, confusion matrix), which is concrete and distinguishable from single-metric siblings such as roc_auc or calibration_curve. The verb is only implied ('compute'), and no sibling is named explicitly, so differentiation is inferential rather than stated.

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?

'USE WHEN: full diagnostic performance set' gives a selection condition that implicitly steers an agent away from single-metric tools. However, it never names the alternative tools (roc_auc, calibration_curve, delong_auc_compare) or states when NOT to use this one, leaving the routing decision to inference.

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