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

Annotations already declare readOnly/idempotent/non-destructive, so the bar is lower; the description earns credit for disclosing the less obvious semantics that a row is one model at its best reasoning-effort configuration, that configurations=true changes the grain of results, and that results are paged. It stops short of describing row contents or ordering/tie-break behavior, which matters given there is no output 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?

Front-loads what the leaderboard is before listing optional modes, and every clause maps to an actual parameter or behavior. It is a single dense, run-on sentence with an ellipsis in the category list, which costs a little readability but wastes no words.

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 six-parameter read-only list tool with no output schema, the description covers the essential grain, filtering, and paging behavior (paging detail lives in the schema's offset field). The notable remaining gap is what fields each returned row contains, which an agent must learn by calling.

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 the schema already documents all six parameters including category and developer examples. The description adds only marginal semantic value beyond that: it clarifies the default grain is best reasoning-effort per model and that configurations flips it, but contributes no syntax or format details the schema lacks. Baseline 3 is correct.

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 resource and verb: a leaderboard of LLMs built from a daily composite of public benchmarks, one row per model. It also disqualifies the confusion with siblings by making clear this is the ranking view (not get_benchmark, get_model, or compare_models), and the modes it supports are enumerated.

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?

Describes the usage modes clearly: default overall rank, rank by a category score, expand to every model × effort configuration, or restrict to open weights / one developer. It never names a sibling as the alternative (e.g. search_models to find a model first, compare_models for head-to-head), so routing guidance is implicit rather than explicit.

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