Ask the user to point at it

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

Annotations already establish readOnlyHint and openWorldHint. Beyond that, the description discloses the genuinely non-obvious behavior: the call blocks on human interaction and consumes a turn, and the return value is selectors ranked by the user's actual clicks. It stops short of saying what happens when the wait expires (empty result, error, partial selection), which matters for a tool whose default wait is three minutes.

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?

Three sentences, front-loaded with the mechanism, then the decision rule, then the cost rationale. Nothing is repeated and no sentence could be dropped without losing guidance.

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?

With no output schema, the description correctly describes the return value (elements plus ranked selectors) and the blocking nature of the call. The remaining gap is timeout behavior — the agent cannot tell whether expiry yields an error, an empty list, or a partial selection — which is a meaningful omission for a tool that pauses execution.

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 a single optional parameter with full schema coverage, including default, minimum, maximum, and a note that the user needs time. The description adds no syntax or strategy guidance for timeoutMs beyond what the schema already says, so the baseline of 3 applies.

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?

"Arms the element picker in the browser and waits for the user to click the elements they mean, then returns them with ranked selectors" names a specific verb chain (arm, wait, return) and a specific resource (user-picked elements with ranked selectors). It is immediately distinguishable from automated siblings like capture_element or collect_selectors because it is explicitly a human-in-the-loop flow.

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 concrete triggering conditions: ambiguous descriptions with multiple matching elements ("the save button", "that broken card") or when the user's target cannot be found. It also names the alternative it replaces (guessing) and quantifies the tradeoff — one turn versus a failed task — which is exactly the when/when-not reasoning an agent needs.

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