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

Annotations already declare readOnlyHint=false, destructiveHint=false and openWorldHint=false, so the bar is lower, yet the description still adds real behavioral context: nothing is sent, the human completes the send, and the tool remains functional even when workspace sending is disabled. It stops short of saying where the draft lands (Drafts folder), whether it can be edited afterwards, or what is returned.

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 short sentences, front-loaded with the core action and no filler. Minor redundancy: 'The person opens and sends it themselves' largely restates 'without sending it', though it does clarify the human-in-the-loop step.

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 7-parameter, 4-required mutation tool with no output schema, the description covers the key safety-relevant behavior (non-sending) but omits return information (e.g. draft identifier), where the draft is stored, and any parameter-level detail. Adequate as a minimum viable definition, incomplete for the tool's actual complexity.

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 description coverage is only 29% (2 of 7 params), so the description is expected to compensate and does not — it mentions no parameter at all, not even the required mailbox/to/subject/body. Notably it never surfaces replyToMessageId's threading behavior, which is the one parameter with meaningful semantic consequence, leaving that 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 ('Write a draft into the mailbox') and the defining constraint 'without sending it', which implicitly separates it from send_gmail_message/send_gmail_reply. It does not, however, name or distinguish itself from the close siblings create_outlook_draft or create_agentmail_draft, so the Gmail-specific routing is left to the tool name.

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?

Usage is only implied: the agent must infer that this is the right tool when it wants to prepare mail for human review rather than deliver it. The line 'Works even when a workspace has switched sending off' is a genuine usage condition but is framed as a capability, not as guidance on when to pick this over send_gmail_message. No exclusions or prerequisites are stated.

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