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

Goes beyond the annotations by disclosing billing semantics (dryRun debits nothing) and atomic pre-flight failure: a 402 refuses 'before any item starts' when the balance cannot cover the total. That all-or-nothing guarantee is exactly the kind of behavior an agent can't get from readOnlyHint/destructiveHint alone. It doesn't address partial failure after start or what the success response contains.

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 what the tool does, then the two conditional behaviors. No filler or restatement of the title.

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?

With no output schema and 5 parameters, the description covers the dryRun/billing contract but says nothing about what a successful batch returns (job IDs? per-item results?), what happens after the pre-flight check passes, or how preset/maxCredits interact with items. Adequate, not complete.

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?

Top-level schema coverage is effectively 0%, so the description must compensate; it does explain dryRun's meaning and implies operation via 'transcribe or burn items'. But maxCredits and preset are never mentioned, and the per-item object shape is left entirely to nested schema descriptions.

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 and resource: 'Quote and run up to 10 transcribe or burn items', which pins down both scope (max 10) and the two operations. Sibling differentiation is implicit via the batching framing rather than naming transcribe or burn_captions directly, so an agent must infer the boundary.

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?

It explains the dryRun path ('returns the quote and does not debit') and the insufficient-balance case, which is real usage guidance. However, it never says when to prefer this over the single-item siblings (transcribe, burn_captions) or what maxCredits/preset overrides are for, leaving the selection decision largely implied.

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