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

Beyond readOnly/idempotent/openWorld annotations, it discloses the decision rule (the winning spelling must be at least 5x commoner), that borderline cases are not silently changed but returned in `undecided` with their options, and that lexicon words are never respelled except for diacritics or broken encoding. These are exactly the non-obvious behaviors that determine whether an agent can trust the output.

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?

Four dense sentences, none wasted, each adding a distinct fact (output shape, use cases, threshold, lexicon exception). The ordering is slightly off for scanning: it leads with the return value rather than with what the tool does and when to reach for it.

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 explained, and the description still characterizes the meaningful parts (changes and `undecided`). With annotations covering safety and full schema coverage on parameters, the definition is sufficient to call the tool correctly; only explicit sibling differentiation 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 description coverage is 100%, including per-value enum documentation for `lang` and `steps`, so the schema already carries parameter meaning. The description adds behavioral context around diacritics decisions and the `undecided` output but never refers to `text`, `lang`, the 20,000-character limit, or the `steps` selection, so it does not go beyond the structured fields.

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 opening line states the result ('cleaned text plus a list of every change') and later sentences make clear it rewrites words only on strong evidence, which together pins down a normalization/fixing operation over Slovenian/Croatian/Serbian text. It is inferable rather than stated as a direct verb+resource, and it never names the sibling checkers it differs from, but the purpose is unambiguous.

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?

It gives concrete usage contexts: text typed without diacritics, speech recognition or OCR output, and analytics. That is a clear when-to-use signal with no explicit exclusions and no named alternative among check_orthography, check_text, or suggest_replacements, so an agent must infer that this tool applies fixes rather than only reporting problems.

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