Perform Action

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

The annotations already carry the safety profile (destructiveHint=true, readOnlyHint=false, idempotentHint=false), and the description adds no behavioral context beyond that: no mention of possible side effects like navigation, whether the action can fail or be rejected, or whether a prior get_app_state is required to obtain valid targets. The description does not contradict the annotations, but it contributes no behavioral insight of its own.

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?

A single front-loaded sentence with no filler; the verb ('Run') and resource ('accessibility actions') lead immediately. It is slightly terse for a tool with a two-variant target union and an optional return_state parameter, but every word earns its place.

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?

Given a mutating tool with no output schema, a complex target union (element vs coordinate), and a large sibling set, one sentence is insufficient. The description does not explain how actions are discovered, what the coordinate target means for an accessibility action, what return_state returns, or what the tool's response looks like. An agent cannot fully determine invocation details from this definition alone.

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 67%, and the schema describes all three parameters with structure, defaults, and enums (action, target variants, return_state). The description adds only the framing that actions are 'advertised by an element,' which loosely connects the action parameter to the target element but does not explain where action strings come from, how to discover valid action names, or what the coordinate target variant is for. It neither strongly compensates for nor ignores the schema.

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 names a specific verb and resource: 'Run one of the accessibility actions advertised by an element.' It is clearly distinguishable from the direct-input siblings (left_click, type, set_value, scroll) because it targets OS/expose-level accessibility actions rather than simulated input. It loses a point because it never explains what an 'accessibility action' is or when an element would advertise one, leaving some inference to the agent.

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?

There is zero guidance on when to use this tool versus the 28 siblings. Notably, set_value and select_text overlap conceptually with accessibility actions, and the description neither names alternatives nor states conditions (e.g., 'prefer this when an element advertises an action'). The agent must guess when to choose perform_action over direct-input tools.

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