draft_message

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

With no annotations provided, the description carries the full burden for behavioral disclosure. It clearly discloses that the tool does NOT send, validates the target is a public channel, renders exactly what would be posted, and returns a one-time confirmation_token. It also explains the token's one-time nature. Minor gap: it doesn't describe error behavior for non-public targets, but core safety-critical behaviors are well covered. No annotation contradiction.

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?

Mostly tight and well-structured. The 'THIS DOES NOT SEND ANYTHING' warning is bolded-in-effect and front-loaded, leading with the critical safety caveat. The Args section is compact. Minor redundancy: 'renders exactly what would be posted' and 'Always show the user this preview' slightly overlap, and 'THIS DOES NOT SEND ANYTHING' plus 'It validates...' creates a touch of repetition. Still, every sentence earns its place and the critical guidance is 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?

Despite having no annotations and 0% schema coverage, the description is fairly complete. It documents the return value (preview + confirmation_token), the workflow with send_message, and validates all three parameters at least at a basic level. An output schema exists which can document the preview/return structure. Gaps: behavior for invalid public-channel target is not described, and there's no mention of rate limits or auth requirements. It's not perfect but is complete enough for a drafting tool with this 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 0% (the JSON schema has no descriptions on properties), so the description must compensate. The description does add some semantic value: channel accepts 'Channel id or #name' and 'Must be a public channel'; text is 'the exact message text to post'; thread_ts is 'Optional thread root timestamp to reply in a thread.' However, the description partially restates what the schema already conveys through names and required-status (thread_ts optional, channel/text required). It adds moderate value but doesn't fully elaborate edge cases or formats beyond what property names suggest.

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?

Purpose is explicit: 'Prepare a message for sending and return a preview + confirmation token.' It clearly states the verb (prepare) + resource (message) and distinguishes from siblings by declaring it is a draft/preview step, distinct from send_message. The negated purpose 'THIS DOES NOT SEND ANYTHING' sharply differentiates it from its sibling send_message.

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?

Excellent usage guidance. It explicitly says 'To actually post, the user must review this preview and then call send_message(confirmation_token=...).' It states the workflow prerequisite, instructs 'Always show the user this preview and get their explicit approval before calling send_message', and references the sibling tool by name. This is exemplary when/when-not guidance.

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