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

Annotations already declare destructiveHint=true and idempotentHint=true, so the safety profile is conveyed structurally. The description adds useful behavioral context via the 'never guess' instruction and the register-first routing for UK/IE companies, but it never explains overwrite semantics — whether omitted fields are preserved, what an existing country/entity_type change does to prior data (the schema notes country is 'refused once there are bookkeeping entries', which the description omits). For a 12-parameter mutation flagged destructive, that is a real gap.

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?

Front-loaded with the action and the field list, followed by the usage rule and the sibling routing, with no padding sentences. The field enumeration is long but functional — it doubles as the list of things the agent should elicit from the user — though it could be trimmed slightly without loss.

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 tool with 12 optional parameters, no output schema and full schema coverage, the description supplies exactly what structured fields don't: when to call it, when to use a sibling instead, and the anti-guessing constraint. Return values are not needed since no output schema exists and the effect is a save. An agent can invoke this correctly from the definition alone.

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%, so every parameter is already documented in detail, including the conditional vat_registered/vat_deregistration_date logic. The description's parameter mentions (sole_trader/ltd, GB/IE/US ISO codes) largely restate schema content rather than adding semantics beyond it. Baseline 3 applies because the schema carries the 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+resource ('Save the user's answers ... to their AccountsOS company') and enumerates the exact attributes covered (entity type, country, registration number, name, activity, currency, year-end, VAT). It also explicitly distinguishes itself from lookup_company and link_company, so the agent can separate it from the read-from-register path without consulting schemas.

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 triggers ('during setup, and when the user corrects one of these details') and an explicit exclusion with the alternative named ('For a UK or Irish limited company with a registration number, call lookup_company and then link_company instead'). It also states the data-collection rule ('Only pass what the user told you; never guess'), which is real routing guidance, not filler.

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