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

Beyond the annotations (readOnlyHint=false, idempotentHint=true, destructiveHint=false) it adds real operational context: the save is free, requires the audiences:write scope, is shared across the organisation, and repeating the call returns the same audience. It further discloses a non-obvious domain default — email audiences include corporate subscribers only under PECR — which the agent could not infer from any structured field.

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?

It is front-loaded with the action and mode split, then usage, then behavior, then compliance caveats — a sensible order with no filler. The PECR/sole-trader sentence is long and slightly tangential to the act of saving, but it carries genuine decision value, so it earns most of its length.

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 save operation with a large nested audience schema and an output schema, the description covers the essentials: modes, auth scope, cost, idempotency, sharing semantics, and where to go next. Only minor gaps remain, such as what happens on a name collision or the significance of brief, which are not covered anywhere else.

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 only 25% at the top level, and the description compensates only partially: it explains audience_id's create-vs-update role and that name labels the saved audience, and mentions that the server fills defaults and returns the effective definition. The brief parameter and the required/optional combinations are left to the schema, and the deep Audience/where-tree semantics are documented entirely in $defs rather than the description.

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 first sentence names a specific verb and resource and separates the two modes explicitly: create ('saves an audience definition under a name') versus update ('updates a saved one when audience_id is given'). An agent can immediately distinguish this from count_audience and list_audiences because the description names those as follow-on lookups rather than saving operations.

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 a concrete trigger ('when the user wants to keep, share or recount a list later') and routes the agent onward to list_audiences to retrieve and count_audience to count. What is missing is an explicit when-not or a note on what to do when saving is unnecessary (e.g. one-off counts), but the positive guidance is unambiguous.

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