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

Annotations already declare readOnlyHint=true and destructiveHint=false, so the safety profile is covered; the description adds genuine behavior beyond that — it only reads stored crawls, and if the latest crawl did not finish it names a laterUnfinishedCrawl rather than silently returning stale data. It also notes definitions arrive in a `note` field, which prepares the agent for the response shape.

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 purpose and the sibling distinction are front-loaded, which is good, but the body is a single ~15-item run-on list of returned data points with no grouping or structure. Since there is no output schema that inventory earns some of its place, yet the density makes it hard to scan.

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?

With no output schema, the description has to carry the return contract and mostly does: status, Site Health and AI Search Health with n and crawl date, error/severity breakdowns, comparison to the prior same-scope crawl, in-progress crawl state, settings and page allowance. Only pagination/volume expectations are unaddressed, which is minor for a single-object read.

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 100% and the schema already says crawlId is 'a finished crawl (from history)' defaulting to 'the latest that ran to the end', so the description largely restates it. The one addition is the laterUnfinishedCrawl fallback pointer for an unfinished newer crawl, which is meaningful but narrow — baseline 3 is right when the schema does the work.

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 names the specific resource ('a client's SEO audit', i.e. a crawl of the client's own site) and states its operation plainly: 'Reads stored crawls only; run_site_audit starts one.' That sentence alone separates it from the mutating sibling, and the enumeration of returned data reinforces what the tool is for.

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 explicitly routes the write case away ('run_site_audit starts one') and states the default selection rule (latest crawl that ran to the end, or crawlId). It does not mention the obvious read alternatives in the sibling list — get_site_audit_history, compare_site_audits, list_site_audit_issues, get_site_audit_page — so the agent still has to infer when this tool is the wrong read.

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