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

Annotations already establish readOnly/openWorld/non-destructive, so the bar is lower, yet the description still adds real traits beyond structured data: per-item error isolation ('each independently ok ... or failed'), ordering guarantee ('in input order'), and question fan-out ('Every question is applied to every item'). It does not disclose volume limits, cost, or latency despite framing itself as high-volume.

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?

Well front-loaded: purpose, then usage discriminator, then question types, then return shape. The middle sentence is dense with stacked parentheticals, and 'Every question is applied to every item' is repeated verbatim between the body and the states schema description, which costs it the top score.

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 return contract itself: JSON Lines, one object per item, input order, success payload keys (choiceAnswers/scoreAnswers/noulAnswers) and failure shape. Combined with 100% schema coverage and annotations covering the safety profile, an agent has everything needed to invoke it correctly.

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 already 100%, so baseline is 3; the description still adds semantics beyond the schema, notably that score rubrics are zero-indexed ('level 0 lowest'), that noul answers are probabilities of 'true' rather than booleans, and how the three question types map to answer keys. That is meaningful extra interpretation on top of an already well-documented schema.

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+resource ('semantic classification and high-volume triage') and concretely enumerates the item types it operates on (search terms, keywords, ad copy, placements, products, audiences). It also contrasts itself against the numeric/rule-based query family that dominates the sibling tool set, so an agent can tell what it is not.

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?

Gives a clear when-to-use condition ('when the answer depends on interpreting language or intent') and a when-not condition ('not on a number you can query or a rule you can filter'), which effectively routes the agent away from the report/query siblings. It stops short of naming a specific alternative tool, so it is strong context rather than fully explicit routing.

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