extract

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

Beyond the readOnlyHint and openWorldHint annotations, the description discloses substantial behavioral detail: boilerplate stripping preserves headings/links/lists/code, output is always markdown text (never a base64 screenshot), results are capped by max_chars, links are deduped and absolute, data: URIs are dropped, and mode=main ignores target subtrees and rejects them. This gives the agent precise expectations for every mode without contradicting annotations.

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 description is compact and efficient, using four sentences to cover purpose, use case, return guarantees, and an exhaustive inline mode reference. It front-loads the core value proposition and immediately supplies the key differentiator (authenticated scraping), with no filler or repetition.

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 seven parameters, no output schema, and no openWorld or readOnly annotations (beyond the ones present), the description fully explains return behavior, mode variations, target scoping, and constraints. It covers all the details an agent needs to choose mode, set limits, and correctly invoke the tool in multi-step workflows, leaving no critical 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?

Although schema coverage is 100%, the description adds meaningful semantics to nearly every parameter: it explains what each mode returns, how target scoping is honored or rejected in main mode, how session names should be made unique to avoid collisions, and how lease tokens are threaded per-call and never read from server env. This is far beyond the schema's one-line descriptions.

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?

Description states a specific action: extracting clean, LLM-ready Markdown from a page or subtree, with boilerplate stripped. It clearly differentiates from siblings like text, html, content, and extract_fields by emphasizing markdown output, subtree targeting, and its role as a drop-in for authenticated-page scraping. The four modes (markdown, links, assets, main) further define exactly what the tool produces.

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 gives a strong usage signal: use it for automated scraping of authenticated pages that stateless tools like Crawl4AI/Firecrawl cannot reach, and to get markdown rather than screenshots. It stops short of explicitly contrasting with internal siblings (e.g., telling when to use text or html instead), so it provides clear context but not exhaustive alternatives.

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