Walk the browser to a spot with Jev

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

Annotations already declare readOnlyHint=false, openWorldHint=true, destructiveHint=false. The description adds significant behavioral context beyond annotations: Jev picks one element/action per step in ~300ms, no page text enters context, Jev never types free text (only names of values are shown), stops before irreversible actions unless allow_irreversible is set, acts on currently selected page, and returns page_id. This is rich behavioral disclosure. It doesn't contradict annotations. Minor gap: doesn't explicitly state that the walk is recorded for follow-up tools in terms of what happens if the page changes, but it does say 'the walk is recorded for them.'

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 a single dense paragraph, but it is front-loaded with the core purpose and then covers key behavioral constraints. Every sentence adds information: purpose, mechanism, follow-up tools, typing behavior, irreversible stops, page selection, return value. It's slightly long but each sentence earns its place. It could be broken into shorter sentences for readability, but it's not bloated. A 4 because it's efficient but a bit dense.

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 the tool's complexity (8 params, nested objects, no output schema), the description is quite complete. It explains the return value (page_id) and how to use it in follow-up calls, which is essential since there's no output schema. It covers the key behavioral constraints (irreversible actions, typing limits, page selection). It doesn't describe error cases or what happens if the goal is unreachable, but the schema's max_steps and timeout_ms parameters imply those limits. The description is complete enough for an agent to call it correctly.

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 the schema already documents all 8 parameters. The description adds meaning beyond the schema: it explains the 'values' parameter semantics (only names reach Jev, not the actual strings), which is critical for privacy/context. It also clarifies 'goal' should describe the end state ('Say what should be visible at the end'). It explains 'url' can be omitted to start from current page. It doesn't add much for max_steps, settle_ms, timeout_ms, tab_url_prefix, allow_irreversible, but the schema already covers those well. Baseline 3 plus the values/goal/url clarifications justify a 4.

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 clearly states the tool's purpose: 'Get the browser to the exact place where you want to look, in one call.' It uses a specific verb ('get the browser to'), names the resource (browser/page), and explains the mechanism (Jev decision model). It distinguishes itself from siblings by explicitly listing follow-up tools like take_snapshot and take_screenshot, and by noting it acts on the currently selected page, differentiating it from navigation tools like navigate_page.

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 provides explicit when-to-use guidance: use this to reach a state in one call, then call list_console_messages, list_network_requests, take_snapshot, or take_screenshot on the same tab. It also gives clear exclusions: 'Stops before orders, payments, sends, deletes, and publishes unless allow_irreversible is set.' This tells the agent when not to use it or what to set. It also explains the alternative of omitting url to start from the current page, and how to pass values for typing, which is a key usage distinction from type_text.

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