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

Far exceeds the annotations: it discloses that the call requires approval and parks for approval, that credit notes draw from a numbered series and inherit the invoice's tax rate, that refunded_outside moves no money, what is refused (draft, cancelled, over-cap, dispute_settlement), and the retry semantics for partial write-offs. Mutations, side effects and failure modes are all surfaced.

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 approval requirement is bracketed at the very front, then type semantics, then units, then refusals, then the next call. Dense but every clause carries a decision-relevant fact; nothing is 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?

For a mutation with no output schema and an approval gate, it covers preconditions, side effects, unit handling and the follow-up read. It does not describe what the parked call returns (pending id, approval state), which leaves a small gap an agent will hit after the call.

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 27% with 11 params, so the description must carry weight and largely does: it defines amount_cents as integer CENTS vs amount as a decimal string with 'never convert units yourself', ties credit_outcome values to balance effects, and explains partial as a deliberate confirmation resent with partial:true. Gaps remain for reason, currency and occurred_at, which are undocumented in both places.

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?

Opens with a specific verb and named resource set: 'Records a credit note, write-off or post-issue discount on an issued invoice.' The three adjustment kinds are enumerated, letting an agent distinguish this from invoice_adjustment_read/list/delete and from invoice_debit_note 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?

Explains selection conditions inside the tool: write_off and discount_post_issue require a 'sent' invoice, credit_outcome chooses reduce_balance vs refunded_outside, and each refusal names the fix. It also routes forward with 'NEXT: invoice_read'. It stops short of explicitly naming sibling alternatives (e.g. invoice_debit_note) as the right call for a different need.

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