desktop_wait

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

Annotations already declare readOnlyHint=true, so the description is expected to add context. It does: it explains that element search targets 'native accessibility nodes,' that element_absent 'requires complete coverage,' and that pixels_stable 'samples the client rectangle, not general application idleness.' These nuances help agents use the right parameters. It omits timeout-failure behavior (error vs. false return) which would be useful, but the read-only annotation covers the safety profile, so this is a mild gap.

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 takes one dense paragraph, front-loading the core purpose before systematically enumerating each condition type and its parameter rules. Every sentence adds operational value—no filler or repetition. The structure flows from general to specific, making it easy to scan despite covering eight distinct modes. It is appropriately concise for the complexity.

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 tool with 9 parameters and 8 condition types, the description covers the essential usage constraints and behavior nuances, including the crucial note that pixels_stable samples the client rectangle and the discovery pointer to desktop_windows. It does not describe what the tool returns on success or timeout (e.g., whether it throws), and it assumes element_id/window_id are obtained elsewhere without detailing how. These are minor omissions given the absence of an output schema, but a complete description would address them.

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 0% (no descriptions in the input schema), so the description carries the full burden. It does so admirably: it maps each condition to its required parameters (e.g., 'element_present/element_absent require window_id plus at least one name/role substring or states filter'), gives value ranges ('timeout (0–10 seconds),' 'stable_for (0.1–10 seconds)'), and clarifies relationships. This exceeds what a bare schema offers, compensating fully for the missing parameter descriptions.

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 states a specific purpose: 'Wait up to timeout for an observed condition.' It enumerates the eight condition types and their parameter requirements, clearly distinguishing this wait/observe tool from its action-oriented siblings like desktop_click or desktop_type. The verb 'wait' plus the resource 'observed condition' is specific, and the coverage of each condition makes the scope 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?

The description gives explicit per-condition parameter guidance: 'window_present/window_absent/window_active require only window_id,' 'text_equals/text_contains require only element_id and text,' and so on. It also advises 'Omit parameters belonging to other condition types' and points to desktop_windows(query=...) for discovering titles, an alternative for that sub-task. However, it does not explicitly contrast with other observation tools like desktop_observe, so agents must infer when this wait-with-timeout tool is more appropriate.

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