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

Annotations only cover the safety profile (readOnly=false, destructive=false, idempotent=false, openWorld=false); the description adds substantial context beyond them, including the preview-scope session requirement, source-grant authorization, immediate start with no confirmation step, provider-side blocking before bytes are sent, and a bounded result with remaining allowance. One mild tension exists: "starts or reuses" reads as dedup/idempotent behavior while idempotentHint=false, though this likely refers to reusing an existing job for the same asset rather than true call idempotency.

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?

Five sentences, each carrying distinct information (accepted inputs, start semantics, policy blocks, result shape), with the core action and input constraint front-loaded. Slightly dense for its length but no filler sentences.

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 mutation tool with an output schema, it covers authorization scope, rejection rules, start timing, and result shape, which is most of what an agent needs. The gaps are the parameter-name mismatch and the absence of an explicit pointer to the sibling used to fetch the finished transcription.

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%, so the description carries the burden, and it does add meaning: sourceAssetId must be a verified asset identifier and arbitrary URLs/paths are rejected, with an optional revision pinning the exact version. However, it names the second parameter "sourceRevision" while the schema calls it "expectedRevision", which can cause a mis-mapped argument, and it does not explain the semantics of the revision beyond it being optional.

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?

"Starts or reuses one exact source audio/video transcription job" gives a specific verb (start/reuse), a specific resource (source transcription job), and a scoping constraint (one exact source) that separates it from retrieval siblings such as get_source_transcription and from the adjacent request_source_image. An agent can select it without opening either 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?

The description supplies real preconditions and exclusions: only a verified sourceAssetId plus optional revision, arbitrary URLs and paths rejected, Google and unverified sources blocked by provider policy, and no separate Apply step. It stops short of explicitly routing the agent to get_source_transcription for polling, though "pending status indicates unfinished work" strongly implies the follow-up.

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