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

Annotations already declare readOnlyHint=true and openWorldHint=true, so the safety profile is covered and the description only needs to add estimation semantics. It does add real value there, claiming architecture-aware KV cache modeling (sliding window, hybrid linear attention, MLA) and engine defaults per GPU class, which tells the agent the numbers are model-specific rather than naive. It still omits what the response looks like, what happens for an unlisted or unknown model, and whether cost is per hour.

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?

Three short sentences with no filler: one on outputs, one on accepted model inputs, one on the estimation methodology. It is front-loaded enough to be scannable, though the lead sentence is a fragment rather than a verb-led statement of purpose, which slightly delays clarity.

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 eight parameters, three enums and no output schema, the description is asked to carry the return-value burden and does so only at the field-name level (memory, count, speed, cost) without units, format, or error behavior. Combined with the double-default rules passed to the schema, an agent has enough to call it, but not enough to interpret the result confidently.

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 every parameter including defaults, enums and the 'auto' option is already documented in the schema; a baseline of 3 applies. The description reinforces the model parameter's accepted formats and hints that kv_cache choices are architecture-dependent, but adds no syntax or constraint detail the schema lacks.

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 states exactly what the tool produces: GPU memory, GPU count, speed and rental cost for running an open LLM, plus the supported model sources. That is more than a restatement of the name, but the opening is a bare noun phrase with no verb ('estimates'), and it never explicitly contrasts itself with the siblings gpu_prices (rental cost overlaps) or models_that_fit (fit-checking overlaps).

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?

Usage is only implied: the agent learns which model identifiers are accepted (the hosted list, any Hugging Face id or link), which suggests when the tool applies, and the 'auto' gpu_count default implies the fit-checking use case. There is no statement of when to prefer it over models_that_fit or gpu_prices, and no prerequisites or exclusions.

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