start_download

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

With no annotations at all, the description carries the full burden and discloses substantial operational behavior: the task is stateful and owned by the current server (shutdown interrupts it), reports survive restarts, thread pools are shared across the whole batch, a service supports at most four batches, no config is mutated, and there is no hidden truncation. It also explains the id vs actual_id distinction and dedup ordering — exactly the kind of context annotations would otherwise need to supply.

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 most important facts — what starts, what is returned, and the follow-up tools — are front-loaded into the first and last sentences, and the middle sentences are information-dense rather than padded. It is somewhat run-on and one clause ('不同时向用户说明两者') is ambiguous, so it is not maximally crisp.

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?

An output schema exists, so return values need not be fully described, and the description still names the key return fields plus lifecycle and concurrency constraints. It is complete enough to invoke correctly, though failure behavior and the precise shape of the returned batch record are left implicit.

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 0%, so the description must compensate, and it largely does: it defines kind=album/photo, explains that ids accepts IDs or links deduped by first occurrence, describes each of album_workers/photo_workers/image_workers (including that the latter two default to download.threading.* config and are scoped differently), and enumerates export_formats values (zip/pdf/long_img). The ids description is entangled with the dedup sentence and does not state accepted formats, preventing a 5.

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 description opens with a specific verb and resource ('立即启动后台批量下载') and states the return values (task_id, log_path, report_path), so the agent knows exactly what the tool produces. It names follow-up tools (get_download_task, cancel_download) and clarifies post_process is unnecessary, but it never distinguishes start_download from close siblings like download_batch, download_album, or plan_download.

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?

It gives clear lifecycle guidance (query with get_download_task, cancel with cancel_download) and notes export_formats removes the need for post_process, which is genuinely useful routing. However, with siblings such as download_batch, download_album, download_photo, plan_download, and retry_download present, it offers no criteria for choosing this tool over them — usage is only implied by 'batch download'.

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