create_draft

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

Annotations only declare the safety profile (readOnlyHint=false, destructiveHint=false, openWorldHint=true); the description adds material behavior beyond that: the draft is never sent, the human must send from Gmail, filesystem reads are restricted to Downloads/Desktop/Documents with hidden folders excluded, and a 24 MB total attachment ceiling. These are exactly the operational constraints an agent cannot recover from structured fields.

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?

Four short sentences, front-loaded with the single most important fact (the draft is not sent) followed by the threading and attachment constraints. No filler, no restatement of the tool name, and every clause carries actionable information.

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 write tool with no output schema and only safety annotations, the definition covers the highest-risk behaviors: non-dispatch, reply threading, attachment scope and size. Minor gaps remain: it does not say what is returned (e.g., a draft ID usable with update_draft) or that the target account must already be connected via connect_account.

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 43%, and the description compensates for just two of the seven parameters: attachments (local file paths, allowed folders, hidden-folder prohibition, 24 MB aggregate limit) and reply_to_id (threading as a reply). The remaining parameters (to, cc, subject, body) get no added meaning, though account is documented in the schema itself.

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 opens with a specific verb and resource ("Creates a draft") and immediately scopes it with "It is NOT sent: the user reviews and sends it from Gmail," which tells an agent precisely what state the mailbox ends up in. It does not, however, name or contrast with the sibling update_draft, so an agent must infer which one applies to modifying an existing draft.

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?

Useful conditional guidance is present for two parameters ("With reply_to_id it goes into that thread as a reply", attachment location/size limits, "Only attach files the user asked for"). But there is no explicit when-to-use-vs-alternative statement against siblings such as update_draft or connect_account, so the routing guidance is implied rather than stated.

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