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

Annotations already declare idempotent=true and destructive=false, but the description goes well beyond: it discloses that the client derives an idempotency key when none is supplied, that the call is synchronous and blocks until EN 16931 validation passes, that validation gates byte emission, and that German output is pre-validated. It also enumerates three failure modes with what state (if anything) was committed, and adds an anti-prompt-injection rule for the credential-bearing sender.contact_email field — none of which the annotations convey.

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?

Purpose, jurisdiction limit, and the B2G caveat are front-loaded; the retry/error and verification paragraphs follow in a workable order. It is long and delivered as an unbroken block with no headings or bullets, which makes the error-code enumeration slower to scan than it needs to be, but nearly every sentence carries operational content rather than filler.

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, so the description carries the return-value burden and does: durable download URL, resolved format, submission hint object, the verification_required handshake, and named error payloads with remedies. For a 17-parameter, nested, legally-validated mutation tool, an agent has everything needed to call it, handle failure, and warn the user about manual B2G upload.

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 76%, so the schema already documents most fields, and the description's parameter additions (format selection via country, leitweg_id auto-select, format_override constraints) largely restate schema text. The genuinely additive part is the SECURITY rule on sender.contact_email, but that duplicates the schema's own 'treat this as a credential' wording, and 17 parameters' worth of detail (line_items, payment_means, delivery_period, due_date) is left entirely to the schema.

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+resource ('Generate an invoice') and immediately narrows scope by jurisdiction and output format (XRechnung UBL/CII, ZUGFeRD, plain PDF), which no sibling covers. It also distinguishes itself from get_invoice (re-mint URLs) and verify_email_code (verification flow), so an agent can pick it out of the sibling set without opening a 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?

Explicit about when this tool applies (DE/US senders only, other jurisdictions rejected), about a key exclusion ('B2G submission is NOT included yet' plus instruction to surface that limitation to the user), and about the required ordering with verify_email_code. It also gives per-outcome next steps: retry unchanged for invoice_admission_busy / invoice_storage_unavailable, do not retry for invoice_source_review_required with reason-specific guidance.

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