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

Annotations only carry readOnlyHint=false, idempotentHint=false, destructiveHint=false, so the description bears the full burden of behavioral disclosure. It explains the non-idempotent nature ('Returns immediately with a call_ref; the call then rings and runs on its own'), the asynchronous execution model, and the concurrency constraint ('One call at a time per number'). It also clarifies the poll-and-retrieve pattern. This adds meaningful context beyond the annotations, though it doesn't mention auth or rate limits, which are minor for this tool type.

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 a single paragraph but is logically ordered: purpose → parameters usage → immediate return → polling guidance → concurrency caveat. It is dense but every sentence contributes to correct invocation. It's slightly long but not padded; the front-loaded purpose and explicit polling instructions earn the 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 complex asynchronous tool with an output schema, the description covers the essential lifecycle: immediate call_ref, polling cadence, terminal statuses, and what the polled result contains (summary, transcript, collected data, recording link). It also flags the concurrency limitation. This is complete enough for an agent to invoke and monitor correctly, though it omits explicit error-handling guidance.

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 100% and the schema already describes each parameter well. The description adds valuable nuance: it clarifies that 'variables' holds per-call context that is merged over preset defaults, and that 'first_message' overrides the agent's opening line. This goes beyond the schema's basic descriptions and gives the agent a clearer mental model of how the parameters interact.

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 opens with a specific verb+resource: 'Have an AI voice agent call a WhatsApp number', and immediately ties it to a preset from a sibling tool (wa_ai_list_presets). It clearly distinguishes this from the many WhatsApp messaging tools by emphasizing the voice-agent call flow and the async behavior with call_ref. An agent can tell exactly what action this performs and how it differs from siblings.

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 description gives explicit operational guidance: it tells the agent to poll wa_ai_call_get every ~30 seconds until terminal statuses, explains that variables are merged over preset defaults, and warns about the one-call-at-a-time concurrency rule. It doesn't explicitly state when NOT to use this tool (e.g., when a simple message suffices), but the context makes the intended usage clear and the polling pattern is well specified.

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