web_crawl

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

With no annotations, the description carries the full burden and does so well: every page goes through the full compliance chain (robots/allow-block list/rate limiting/same-host concurrency), a single page failure does not abort the crawl, static asset links are skipped automatically. It also discloses non-obvious side effects: explicit headers/cookies/proxy bypass the shared cache, and when deadline_ms is exhausted already-fetched pages are still returned while queued subtasks return timeout.

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 purpose and mode split are front-loaded, then a structured Args list follows — the bulk is necessary given zero schema descriptions. The trailing ctx entry is framework plumbing that adds little for an agent, a small amount of dead weight.

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 15-parameter crawler with no annotations, the description covers mode selection, failure isolation, caching, deadlines, and compliance — enough to call it correctly. Return values are covered by the existing output schema, but auth/session prerequisites and concurrency limits are only gestured at.

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 Args block is doing all the work — and it documents 14 of 15 parameters with real semantics: depth range 0–3, limit bounded by crawl_max_pages, next_selector suppressing depth, max_pages default 5, session applying to all pages. It loses a point because a few details (e.g. timeout/proxy fallback nuances vs. global config) remain terse, but the compensation for a 15-param, 0%-covered schema is substantial.

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 opening states a specific verb and resource: discover in-site links from a start page and batch crawl them, or page through via "next page". It further splits the tool into two named modes (link mode / pagination mode) so the agent understands the tool's two distinct behaviors. It only references a sibling indirectly ("link_policy: 同 web_fetch") rather than explicitly contrasting scope with web_fetch/web_batch, so it stops short of a 5.

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?

"适合把一个栏目的前 N 篇读进来" gives a concrete selection scenario, and the two-mode block states which parameter selects which mode (next_selector overriding depth). There is no explicit "when not to use this / use web_fetch for a single page instead" exclusion, which keeps it at 4 rather than 5.

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