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

Annotations already indicate read-only, idempotent, nondestructive behavior, so the description adds meaningful behavioral context beyond that: fresh runs report PROCESSING, wait: true blocks until terminal, non-terminal runs should be re-polled, and terminal statuses are PROCESSED, FAILED, CANCELLED. It even clarifies aggregation semantics of accuracy metrics across items, which is not visible from annotations or the schema.

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 description is dense and front-loaded with the core purpose and its unique value, and every sentence adds operational value. It is fairly long and packed into one paragraph, but it earns its length by covering async semantics, terminal statuses, and metric shapes. Slight structural improvement would come from breaking it into bulleted sections.

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?

Given the async nature of evaluations, the presence of an output schema, and the complexity of different metric shapes per resource type, the description is complete. It tells the agent how to handle polling, what statuses to expect, and what metrics are returned for extractors, classifiers, and splitters. Nothing needed to invoke the tool correctly is missing.

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%, so the schema already documents all parameters thoroughly. The description adds important context for 'wait' behavior and the relationship between evaluationRunId and run_evaluation, plus metric names per resource type. This elevates it slightly above the baseline but does not need to carry the parameter documentation burden.

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 verb and resource: 'Get an evaluation run's status and accuracy metrics.' It also explicitly differentiates itself from get-batch tools by saying it is the ONLY tool that returns these metrics, making its purpose unambiguous among a large group of sibling tools.

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?

Provides explicit guidance for the async job scenario: pass wait: true to block, and call again with wait: true rather than re-submitting if still non-terminal. It also names the alternative category (get-batch tool) and distinguishes the evaluation-run use case, so the agent knows when to use this tool vs alternatives.

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