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

The description adds substantial context beyond the annotations: the async job pattern (returns a Job ID, then call wait_for_jobs), per-item latency of 1.5-3 minutes for video and ~20 seconds for images, that peaks tracking is slower, and credit costs of roughly 19-76 per video item. The annotations only declare readOnlyHint=false, openWorldHint=true and idempotentHint=false, so the credit and latency disclosures are the real value. However, it does not explain why readOnlyHint is false (credits/job creation) or what happens on failure or partial completion, leaving some behavioral gaps.

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?

Front-loaded with the payload description and followed by operational guidance, with no filler sentences. The first sentence is a long dense enumeration that could be broken into bullets, but every element listed is decision-relevant for an agent choosing this tool.

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?

There is no output schema, and the description compensates by explaining that the call returns a Job ID, that results arrive via wait_for_jobs, and that returned times are in the asset's own seconds. Combined with cost, latency, batching and flag guidance, an agent has everything needed to invoke and follow up 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 75% and the description meaningfully enriches the two flags: peaks returns the measured frame of each action's hit and tracks actors (slower), faces returns who is visible when, lip visibility and the active speaker. It also clarifies that times are the asset's own seconds and that ranges should be 5-60 s. projectId is left to the schema, which documents it adequately.

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 ('Read what is in ranges of the project's videos and in its images') and enumerates the readings produced: people/objects with ids, events and timing, transcript and soundbites, on-screen text, usability, loudness/music/beats, and text-safe areas. This is clearly distinct from search_footage (which locates ranges) and wait_for_jobs (which retrieves results).

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: 'Find the ranges first (search_footage), then inspect only those', with the typical range size of 5-60 s and the instruction to batch every range into one call. It also states when to enable peaks (cutting on an impact) and faces (who is visible, lips, active speaker).

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