Run UI task

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

Annotations only provide readOnlyHint=false and openWorldHint=true, which are minimal. The description goes beyond these by disclosing a clear side effect ('The session is left on whatever page the agent ended on'), the report structure (success flag, summary, findings with severity/where/evidence), and that findings originate from both the agent and the harness (console errors, failed requests, dialogs). This adds substantial behavioral context, though it could mention failure handling or resource implications more explicitly.

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 two paragraphs, with the core purpose and usage guidance front-loaded. It efficiently uses examples and explains the return format and side effects. While it is a bit lengthy, every sentence adds value: purpose, examples, return structure, and session side effect. It is well-structured and free of filler.

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 delegation tool with five parameters and an output schema, the description covers the essential operational context: what it does, when to use it, what it returns, and side effects. The presence of an output schema covers return details, and the description mentions the session side effect for follow-up. It is slightly incomplete regarding error/timeout behavior, but adequate given the tool's complexity.

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 100%, so all five parameters (sessionId, instruction, model, maxSteps, expectation) already have descriptions. The tool description does not add parameter-specific semantics beyond giving usage examples for 'instruction' and mentioning the report structure, which is not parameter-relevant. Since the schema carries full parameter meaning, the baseline is 3; the description does not exceed it.

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 verb ('Delegates a multi-step UI task to a fast built-in browser agent') and resource (UI task), with concrete examples. It clearly distinguishes from siblings by contrasting with 'driving each click yourself' and listing browser_* tools as the step-by-step alternative. An agent can immediately understand the tool's unique role.

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 explicitly says 'Use it for goal-shaped work' and provides two concrete examples ('log in as demo@example.com and check the dashboard loads', 'walk the checkout flow and report anything broken') versus driving clicks. It implicitly tells when not to use it (for step-by-step control) via the contrast with browser_* tools, giving clear selection guidance.

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