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

Annotations already declare the safety profile (not read-only, not destructive, not idempotent), so the bar is lower, and the description still adds meaningful behavior: steps runs up to 20 actions, stops at the first error, and returns one final diff; the default post-action view is a diff. It omits auth/secret handling nuance and any error-shape detail, keeping it below a 5.

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?

Nearly every clause carries operational information — nothing is filler. Structure is weaker: it is one long run-on string with no line breaks, and it front-loads an upload edge case before the core 'what/when' framing, forcing the agent to scan past it to find the action list.

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 an 8-parameter dispatch tool the coverage is decent but has holes: cookies, storage, and emulate are listed with zero explanation of what they do or expect, and close_session appears in the enum while a sibling close_session tool also exists — an ambiguity the description never resolves. With no output schema, the description should say more about what cookies/storage return.

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 already 88%, so baseline would be 3, but the description earns more by binding parameters to each kind: type takes ref+value, press takes value=key, select takes ref+values, drag takes ref as source with args.to or args.to_target, and upload specifies the exact args.files=[{name,text|base64}] shape. That is genuinely additive for the free-form nested args object.

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 concrete verb+resource ('Act on the page') and then enumerates the full action vocabulary (click, type, drag, upload, dialog, steps), which makes the tool's scope unambiguous. It doesn't explicitly differentiate itself from siblings like evaluate, navigate, or extract, so an agent must infer that act is the interaction layer versus those tools.

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 real situational guidance embedded — 'for yes-flows first value=policy... then repeat the click', 'for a drop area without a ref use args.to_target', 'draggable elements are marked in the snapshot'. But there is no explicit statement of when to choose act over evaluate/navigate/snapshot, and no exclusions or preconditions (e.g. session must exist first).

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