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

Annotations only give readOnlyHint=false and destructiveHint=false; the description goes far beyond them by disclosing the cost model ('MOST EXPENSIVE: per second, on completion, purchased credits only'), the async pattern and polling target ('poll get_task'), and a free quoting path ('dry_run:true (free), confirm with the user'). This is exactly the operational context an agent needs before spending credits.

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 cost warning is front-loaded, followed by a tightly structured per-model routing list, and almost no sentence is wasted. The telegraphic style ('180/240 per second of the clip (3-30s), silent') is dense to the point of being cryptic in places, which costs a little readability but not correctness.

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 22-parameter, multi-model, no-output-schema tool, the description covers pricing, model selection, async polling, quoting, and entity attachment, which is close to complete. The main gap is the return value: it implies a task id via 'poll get_task' but never states what the call returns or how to identify the resulting generation.

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?

With 73% schema coverage the baseline is 3, but the description adds genuine meaning beyond the schema: model-specific constraints such as '3 reference_image_ids (6 with no clip)', duration and resolution behavior per model, the draft/finalize_draft_id workflow, and how entities attach ('@slug or entity_ids'). It leaves a few parameters (name, aspect_ratio, extend_direction) to the schema, which is acceptable.

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 opening 'Makes a video' is thin and nearly restates the title, but the body makes the resource and scope unmistakable: it generates video clips from stills, references, or text across seven named models. It is clearly distinguishable from generate_image and add_video by resource, though the differentiation is implied rather than stated.

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?

Strong routing guidance: each job type ('Still to life', 'Cheapest with sound', 'Copy a dance', 'Edit a clip in words') is mapped to a specific model with credit costs, so an agent can pick a model without guessing. It also prescribes a workflow (quote with dry_run, confirm with the user). What is missing is any when-not-to-use guidance versus sibling tools such as add_video or expand_image.

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