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

Annotations already establish readOnly, idempotent and non-destructive behavior, so the description's job is added context — and it delivers the important one: at most 2000 rows are parsed, plus the exact shape of what comes back (columns, rows, and raw CSV text as Joist produced it). It does not say what happens when the month exceeds 2000 rows (silent truncation vs. error), which is the main behavioral 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?

The column list is stated twice in near-identical form ("number, client, subtotal, tax, total, dates, payments" and then again with per-tax-name columns and payment-received detail), which makes the middle of the description redundant. The key scoping facts (one month, date issued, 2000-row cap, no line items) are present but not front-loaded.

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 correctly takes on the burden of describing the return value, and it does so concretely. What is missing is edge-case behavior — truncation handling, empty-month results, and whether the returned CSV text is identical to the UI download for partial pages — but for a two-parameter read tool this is close to 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?

Schema coverage is only 50%, and the description compensates meaningfully: it explains that `type` selects the estimates or invoices export and, crucially, that `month` is interpreted by date issued rather than date created — semantics the pattern-only schema does not convey. The tax-column behavior (one column per tax name) and the invoices-only payment column also add real interpretation value.

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 a specific verb and resource — a monthly CSV export of estimates or invoices, mirroring the Export button on Joist's Estimates/Invoices page — and enumerates the exact columns produced. It also distinguishes itself from the record-level siblings (get_estimate, get_invoice), which return line items this export omits.

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 clearly scopes use to one month, by date issued, and explicitly routes the agent elsewhere for line-item detail ("Line items are not in the export; use get_estimate or get_invoice"). It does not compare against other bulk-read siblings such as list_estimates, list_invoices or get_report, so an agent choosing between bulk-list and export still has to infer.

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