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: it discloses cost behavior (no model calls, no wallet debit), latency (under a second), the error contract (audience_size=null plus spec_errors/spec_error_details with the same stable codes a run would reject with at 422, before any charge), and the return shape including audience_note/incidence semantics. It also pre-empts a likely agent misread — that 'small' means blocked — which is exactly the non-obvious behavior an agent 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?

Front-loaded with the free/sub-second hook, then the small-audience caveat, then the returns block and error contract — a sensible order. It is long and partly duplicates the targeting schema's prose about small audiences, but nearly every sentence carries decision-relevant information, so the bulk is defensible.

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 supplies an explicit Returns block (audience_size, narrowed_by, narrowing_filters, preset_key, optional by_dimension) and the full failure surface with error codes. Combined with nested-object targeting params, an agent has everything needed to call it, interpret results, and act on failures without guessing.

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 the baseline is 3, but the description adds real meaning: the preset_key/demographic_preset aliasing rule (send either; both must name the same market), and the critical interest-tag guidance that a ~200-topic taxonomy means a single tag is a narrow slice and several tags are OR-matched. It does not explain sample_size or include_by_dimension's cost/effect beyond the schema text.

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 — preview how broad an audience a `custom_audience` filter describes — and immediately scopes it as a free, sub-second validation step distinct from a paid run. An agent can distinguish it from `crowd_sample`, `preview_cost` and the run tools without opening any 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?

Explicit timing is given: 'Call it before a run that uses custom_audience to check the spec is valid and to see how tight the cut is,' plus a strong when-not: never abandon a run because the band returns 'small'. It routes to sibling concepts ('the run tools', fix the spec here before a 422) but never names `preview_cost` or `list_demographic_presets` as alternatives, leaving some inference about which free pre-flight check to pick.

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