create_dataset

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

Annotations already declare readOnlyHint=false, destructiveHint=false, and openWorldHint=false. The description adds one genuinely useful behavioral fact beyond that: name collisions are rejected rather than overwritten, so the agent understands this is a non-destructive create-or-fail operation. It adds nothing further about validation rules, size limits, or whether text and messages may be mixed within one example.

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?

Two short sentences, zero redundancy, with the core action front-loaded and the constraint second. Nothing could be removed without losing information.

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?

An output schema exists, so return values need not be described, and annotations already cover the safety profile. Still, for a mutation tool with a nested example union, the description omits whether text and messages can be mixed, what happens when both are null in an example, and any permission or size expectations — gaps a caller would hit.

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 0%, so the definition must compensate. The phrase 'from text or chat examples' meaningfully explains the two mutually alternative shapes of the Example object (text vs. messages) that the schema only encodes structurally. However, the `dataset` name pattern, the per-example length limits, and the distinction between raw text rows and role/content message rows are left entirely to the schema, so compensation is partial.

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 plus resource plus output format: 'Create a new JSONL dataset from text or chat examples.' An agent can immediately tell this is a write/create operation producing a dataset, distinct from inspect_dataset or create_tensor_asset. It stops short of explicitly naming a sibling or contrasting scopes, so a 4 rather than 5.

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?

The only conditional content is 'Existing names are rejected,' which is a failure rule, not when-to-use guidance. There is no statement of prerequisites, no naming of an alternative tool (e.g., inspect_dataset for reading, create_tensor_asset for tensor data), and no guidance on when a caller should prefer chat-style vs text-style examples.

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