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

Annotations cover only readOnly/openWorld, so the description must add the substance, and it does: it discloses the return shape (a short chat/form message plus a full letter), the PDF side-effect available at refund_report_url, the reject rule that confirmation number and guest name are not accepted and render as placeholders, and the enrichment behavior when tracking_id is supplied. These are exactly the traits an agent cannot infer from annotations or schema.

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 dense sentences, front-loaded with the action and the four kinds before the operational caveats. Every clause carries information, though the kind-specific parentheticals make the opening sentence long enough that an agent must parse carefully.

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 25-parameter, no-output-schema tool, the description covers the important gaps: it explains the four modes, the price_guarantee preconditions, the placeholder rule for identity fields, and the response contents so the missing output schema is less damaging. It is slightly thin on the minor parameters and does not spell out the full structure of the two returned texts.

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?

With 72% schema coverage the schema already carries a lot, yet the description adds real cross-field semantics the schema does not: booked_on/paid_total/lower_total/lower_price_site are grouped as price_guarantee requirements, new_check_in may be omitted to ask which dates are possible, and tracking_id injects recorded Agoda prices into specific letter types. Remaining parameters (city, room_type, adults, cancellation_policy, breakfast) are left unexplained in the description.

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?

Names a specific action (write the text of a request) and enumerates the four exact kinds supported, with a one-clause gloss for each (claim for the price difference, date_change, goodwill_refund for non-refundable bookings, checkin_request citing a lower price). An agent can distinguish this from siblings like get_refund_guide or check_refund_claim without opening any schema.

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 clear per-kind selection guidance, including the critical precondition that only goodwill_refund applies to non-refundable bookings and that price_guarantee is a claim against the booking site rather than the hotel. It never states when NOT to use this tool or points to a sibling alternative, so it stops short of a 5.

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