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

Annotations already declare readOnlyHint=false, openWorldHint=true, destructiveHint=false and idempotentHint=false, so the mutation profile is known; the description adds non-obvious behavior it could not infer: deduplication when an audit is already running, the response signal 'created=false', and the fact that desktop PageSpeed always runs while mobile is opt-in. It does not mention expected duration, external fetch cost, or quota limits, which would matter for an open-world crawl trigger.

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?

Three short sentences, front-loaded with what the tool does, then the parameter nuance, then the follow-up tool. Every sentence carries distinct information; nothing is redundant padding.

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?

There is no output schema, and the description partially compensates by disclosing the 'created false' return signal and the hand-off to watch_audit. It stops short of naming the identifying field of the returned audit that watch_audit would consume, which is the one gap for a 2-parameter async start tool.

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 both parameters are already documented, including 'from list_sites' and the default of includeMobile. The description only restates this ('Set includeMobile to add mobile PageSpeed. Desktop PageSpeed always runs'), adding no syntax or format detail beyond the schema, so the baseline 3 applies.

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?

States a specific verb and resource ('Starts an SEO Audit of a site's saved domain') plus the concrete engine used ('desktop PageSpeed'), and anchors the behavior to a known UI action ('Run audit' in Keytomic). It is clearly distinguishable from watch_audit, list_audits, and the other start_* siblings.

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?

Gives explicit next-step routing ('Follow it with watch_audit, which streams its progress until it ends') and handles the already-running case by returning the existing audit. What is missing is guidance on the alternative flow after completion (whether to poll get_audit or go to list_audit_findings), so it stops short of full when/when-not/alternatives coverage.

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