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

Annotations already declare this is a non-read-only, non-destructive, closed-world operation, and the description adds real value on top: async job semantics, typical completion time, and a per-second credit cost. That said, the cost model (half a credit per second, rounded up) and any failure behavior are stated only in passing, and the openWorldHint=false annotation sits awkwardly with fetching external https photo/audio URLs — not a contradiction, but unexplained.

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 dense sentences with the artifact and its constraints front-loaded, then the async return contract, then cost. No filler and nothing repeated from the annotations or schema.

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, the description correctly supplies the return contract (job id, poll video_status) and the latency and cost expectations that determine whether an agent should call it. Limits on scene count and per-scene duration live only in the schema, so the prose picture is slightly incomplete but still sufficient to call correctly.

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 low at 33%, so the description has to compensate, and it does for the main surface: it explains what a scene contains (a filled template or https photo, shown for some seconds with slow zoom or pan), that transitions join scenes, and that music is optional. It stops short of enumerating the motion enum values, the 1–30 second bound, the 20-scene cap, or what "modifications" slot values mean beyond the schema's "as in render_image".

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 artifact ("Make an MP4 (up to 90 seconds) from scenes") plus the composition rules — templates or https photos, per-scene seconds, zoom/pan motion, fades. An agent can tell this produces a video, not a still image or document, without opening the schema.

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?

Gives an explicit workflow for the async result: the call returns a job id at once and the agent should call video_status with it until status is done (typically 15–60s). It does not contrast this with render_image or explain when a single image render is preferable, so there is context but no stated exclusions.

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