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

Annotations already declare readOnlyHint, idempotentHint and destructiveHint=false; the description's 'Read-only, spends no credits' largely restates that. It does add real signal beyond the annotations — the response shape and status values vary by kind, and the character_reference case is spelled out as `{ ready: false }` or `{ ready: true, referenceUri, previewUrl }`. It still doesn't enumerate what the other kinds return, or what to do on failure/expiry.

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?

Purpose and the do-not-call conditions are front-loaded, which is right, but the long inline mapping of thirteen kinds to tool names is dense prose that would read better as a list, and it pushes the character_reference and seriesId guidance into a trailing run-on. Size is justified by the breadth of kinds, but the structure is not.

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 eight parameters, four of them required, and no output schema, the description carries the full burden of describing return values — and it only says the shape 'depends on kind' while detailing a single kind. For a polling tool, the absence of typical status values or terminal-state behavior leaves a real gap.

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 88%, so the baseline is 3, and the description earns above that by explaining that the returned identifier may be named jobId, job_id or taskId, that `kind` names the originating tool, and that seriesId is the flowId of the Flow Studio flow obtainable from list_flows. The analytics-only parameters (context, llm_model, conversation_id) are left entirely to the schema, which is acceptable at this coverage level.

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 — check one async generation job and retrieve its result URLs once finished — and makes the polling semantics explicit. It goes further by naming exactly which sibling tools produce which `kind`, and by routing `run_flow` and `generate_episode` elsewhere, so an agent can distinguish it from every relevant alternative.

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 positive conditions (after any tool that returns a job identifier), negative conditions (don't call when the id is missing, or when generation already returned a synchronous result), and concrete alternatives (get_flow_run for run_flow, the 'status' action for generate_episode). The kind-to-tool table removes any ambiguity about whether this tool applies.

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