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

The description discloses credit consumption, page-pooled pricing, caching behavior, and the always-present summary field. This goes well beyond the sparse annotations and aligns with readOnlyHint=false, since categorization incurs credits on first use.

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 description is front-loaded with the primary action and then adds cost/caching and response guidance. There is slight redundancy between 'Returns cached categories with no charge' and 'free on every re-fetch after,' but the paragraph is compact and focused.

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?

For a single-parameter tool with an output schema and annotations, the description covers prerequisites, cost behavior, caching, and response interpretation ('summary' field). An agent has enough information to call it correctly without external context.

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?

There is only one parameter and the input schema already describes document_id with 100% coverage, including its source. The description's 'previously processed' reinforces the prerequisite without adding significant new semantic detail.

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 names a specific verb and resource ('Run AI transaction categorization on a previously processed document') and states the deliverable ('return its category mappings'). It clearly distinguishes this tool from listing, conversion, and comparison siblings by focusing on categorization of an already-processed document.

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 phrase 'previously processed document' and the parameter description 'from convert_statement or list_statements' give clear prerequisite context for when to call it. It does not explicitly name alternatives or exclude cases like raw/unprocessed documents, but the intended usage is evident.

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