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

With no annotations, the description shoulders the behavioral burden and does it thoroughly: it reveals that the trace is diagnostic rather than replayable, that payloads over 2000 chars are truncated, and that appending stops after 50,000 entries. It also explains why CDP payloads carry session-scoped ids and why a screenshot alone returns megabytes of base64. This is exemplary behavioral disclosure for a recording tool.

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 about 70 words in four sentences, with the main action and scope first, followed by the diagnostic caveat and limits. Every sentence contributes a distinct fact: what is recorded, scope, why not replayable, and hard limits. It is dense but not bloated.

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?

The description covers the recording's lifecycle: what is captured, how long it runs, the stop condition, and hard limits. Gaps are minor: it doesn't state what the start call returns (e.g., a recording ID), and the stop-tool reference appears to be a typo ('ios_safari_record_stop' vs. the actual sibling 'ios_actions_record_stop'). Still, for a start-recording tool it is largely complete.

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?

The schema already documents both parameters with 100% coverage, and the description adds little about udid or label beyond the schema's own text ('Short name to find this recording by later'). The description's detail is about trace behavior, not parameter meaning. Baseline 3 is appropriate given the schema carries the full load.

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 opens with a specific verb and resource: 'Start recording every Web Inspector / CDP command and event for a device — what was sent, what came back, and how long each took.' It also distinguishes itself from replayable scripts by saying 'use flow recording (which captures selectors) when you want playback,' which separates it from replay-related siblings. The only wrinkle is naming the stop tool as 'ios_safari_record_stop' when the sibling list shows 'ios_actions_record_stop,' but the core purpose is unambiguous.

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?

It provides an explicit alternative: 'use flow recording (which captures selectors) when you want playback,' so an agent knows when not to choose this. It also defines the scope ('Spans all Safari tool calls until ... stop') and characterizes the result as a diagnostic trace. However, it references 'ios_safari_record_stop' as the terminator while the actual sibling is 'ios_actions_record_stop,' which is a minor but real guidance error.

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