text_classify

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

With no annotations, the description carries the full burden and does so richly: input limits (6,000 chars), label count bounds (2–30), JSON-schema-enforced outputs guaranteeing answers come from the caller's labels, multi_label capping at 5, prompt-injection resistance, language coverage, model identity, per-call cost ($0.002 USDC), and p99 latency. This is exactly the behavioral disclosure a caller needs.

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?

The purpose and use cases are front-loaded and the paragraph is dense with no filler, but it is a single long block that mixes usage examples, I/O semantics, and operational metadata. A little structure would improve scannability without losing content.

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?

No output schema exists, so the description takes responsibility for return values — best label with a 0–1 confidence and one-sentence reason, or all applicable labels under multi_label, always constrained to the caller's own labels. Combined with size, cost, and latency limits, nothing needed to invoke it correctly is missing.

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 a 3 would be defensible, but the description adds meaning beyond the schema: it explains optional per-label descriptions, the multi_label switch semantics (best label + confidence + reason vs. every applicable label), and that label adherence is enforced rather than suggested. It does not clarify the 'instructions' parameter's role, which holds it below 5.

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?

Starts with a clear verb+resource ('Text classification with your own labels') and immediately enumerates concrete agent use cases — route support tickets, qualify leads, tag feedback, sort emails, moderate comments — that make it unmistakably distinct from siblings like language_detect or news_sentiment. An agent can select it without opening the schema.

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 enumerated use cases give strong context for when to reach for this tool, and the labels/multi_label framing clarifies the two operating modes. However, it never explicitly names an alternative sibling (e.g., language_detect for raw language ID) or states when NOT to use it, so it stops short of the 5 bar.

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