Export a dataset

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

Annotations declare readOnlyHint=false, destructiveHint=false, idempotentHint=true, so the safety profile is already carried; the description adds genuinely new context: files are written to a folder on the server machine, that option is unavailable when hosted, and only a summary is returned in that case. It does not cover the returned-payload-vs-files tradeoff beyond that or any size/perf limits besides the 100-image cap.

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-loaded with the core purpose, then format/selection/output details in a compact block. Some phrasing duplicates schema descriptions (classes ordering, output_dir folder semantics), but each clause still carries operational information rather than filler.

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 needn't be spelled out, and the description still notes that only a summary comes back when writing to disk. The one gap is that with zero required parameters, the behavior when neither category nor image_ids is given (error? export all?) is never stated.

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 baseline is 3, but the description adds meaning beyond the schema: how classes fixes numbering and that other labels follow first-seen order, what each format produces ('one COCO file for all images', 'a label file per image, named after it'), and the hosted-server restriction on output_dir. Only the all-optional default case is left unexplained.

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 with its defining scope: 'Export several images as ONE training dataset, with one class list so ids mean the same everywhere.' The three output formats are named, and the distinction from export_annotations is implicit in the resource (multi-image dataset vs annotations) and made concrete by the format list.

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 clear selection guidance ('by category (case-insensitive) or image_ids (up to 100)') and a situational recommendation ('output_dir ... suits large datasets such as nuclei'). It does not name a sibling alternative or state what happens when neither category nor image_ids is supplied, so it stops short of full when/when-not coverage.

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