Documarium

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

Annotations already declare destructiveHint=true, idempotentHint=false, and readOnlyHint=false, so the safety profile is covered. The description adds that the result is a numbered version, which is useful versioning context, but it omits atomicity, permissions, and what happens to prior versions.

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 definition is a single front-loaded sentence with no filler. Its brevity is appropriate, though 'Keep' is slightly colloquial and less precise than 'apply'.

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 destructive mutation with optional site, draft_id, and proposal_id selectors plus an operations array, the description is too thin. It does not explain the relationship between draft_id and proposal_id, atomicity, or the expected versioning outcome, relying heavily on schema and annotations.

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 100%, and each parameter already has detailed descriptions, including draft_id, proposal_id, operations, and all-at-once application. The description adds no parameter-specific meaning beyond this strong schema baseline.

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 action (keep a staged set of changes) and outcome (saved as a numbered version of the site). It communicates a commit-like operation, though it does not explicitly distinguish itself from siblings such as propose_edit, publish, or rollback.

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 no explicit when-to-use or when-not-to-use guidance, no prerequisites, and no named alternatives. At best, the phrase 'staged set of changes' implies it is used after a staging step, but the agent must infer this.

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

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

The description adds meaningful context beyond the annotations by saying that keeping a preview confirms ownership or permission and that this is recorded. However, with destructiveHint=true declared, it does not explain what happens to the preview, whether the action is irreversible, or what side effects occur beyond the ownership record.

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 short and front-loads the main action before qualifying behavior. Every sentence contributes useful information, though it could be slightly tighter.

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 two-parameter tool with no output schema, the description covers the main action and ownership confirmation but omits what the tool returns on success, whether the preview remains available, and the concrete impact implied by destructiveHint=true. It is adequate but leaves notable behavioral gaps.

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 100%, so both parameters are already documented in the schema. The description reinforces that omitting subdomain yields free name suggestions, but it adds little meaning beyond the schema's own parameter 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?

The description states a specific action: keeping a preview as a site and choosing the web address name. It references create_preview via the preview concept, so the agent can distinguish it from preview creation, but it does not explicitly contrast with publish or site_status.

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?

Usage is implied by 'Keep a preview as a site of your own' and the schema note that draft_id comes from create_preview. It does not state when to choose this over publish or other site-management siblings, nor does it give exclusions.

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

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

Annotations already establish this is a non-read-only, non-destructive, non-idempotent operation, so the bar is lower, and the description still adds real behavior: the site is temporary and unclaimed until kept, a link is returned, the response includes per-page items to check, and pages containing notes such as [GAP], [?] or TODO are refused. It does not cover auth requirements, expiry window, or viewer access for the preview link, which are the remaining gaps.

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 first sentence is well front-loaded, but the remaining text is a run-on stream mixing content policy, legal-field rules, translation paths and page granularity with no paragraph or list structure. Most sentences carry a genuine constraint, yet phrasing like 'exactly as it happened' and 'a page missing one of them still reads complete without it' costs clarity. It needs restructuring more than trimming.

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 seven-parameter, nested-schema tool with an output schema, the description covers the parts structured fields cannot: content acceptance rules, source-selection behavior, legal-page constraints, and translation layout. Return values need not be explained because an output schema exists. Missing are operational details such as how long the preview lives, who can access the link, and any size or rate limits.

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 100%, so baseline is 3 and the schema already explains pages, source.kind, logo_url and languages. The description adds some meaning, notably that languages drives all page text and headings and that translated copies appear at /<code>/ paths, but most parameter prose restates the schema's own wording rather than extending it.

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 first sentence gives a concrete verb and resource: turns markdown pages into a live preview site and returns a link. That is enough to tell it apart from siblings like read_page, list_pages or propose_edit. It never names publish or claim_draft explicitly, so the boundary between previewing and keeping/publishing a site is only implied by 'temporary and belongs to no one until somebody keeps it'.

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?

Usage is implied rather than stated: the preview is temporary and the result tells you what to check before sharing the link, which suggests it precedes publish, but no alternative is named and no when-not condition is given. The bulk of the text is content-entry requirements (page granularity, no TODOs, legal fields) rather than tool-selection guidance. An agent would infer the workflow but not be routed explicitly.

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

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

Annotations already declare readOnlyHint, idempotentHint, destructiveHint=false, and openWorldHint=false, so safety behavior is covered. The description adds a useful behavioral detail: it includes passages that express a fact differently, implying paraphrase matching. However, it does not describe result scope, limits, ranking, or whether the search is exhaustive.

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 a single efficient sentence with no filler. It is front-loaded with the core action, though the wording 'states a fact' is slightly ambiguous. It avoids unnecessary 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?

For a search-like tool with three parameters and no output schema, the description is thin. It does not explain what a returned 'passage' looks like, whether results are ranked or paginated, or how site and draft_id affect scope. The agent must rely almost entirely on the schema and name.

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 100%, so the schema already explains query, site, and draft_id. The description adds no meaning beyond the schema for any parameter. Baseline 3 is appropriate because the schema carries the parameter semantics.

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 says it finds passages stating a fact, including paraphrased passages, which gives a rough sense of a semantic reference search. However, it does not clarify what a 'reference' is here or distinguish the tool from siblings like read_page or history. The purpose is implied but not precise.

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?

There is no guidance on when to use this tool versus alternatives such as read_page, list_pages, or history. It also omits any when-not conditions or prerequisites. The agent is left to infer usage from the name and schema alone.

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

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

Annotations already declare readOnlyHint=true, idempotentHint=true, destructiveHint=false, and openWorldHint=false, so the safety profile is fully covered by structured data. The description adds only the substantive behavior that the same tool serves two modes, listing and comparing, which is genuinely useful context. It does not disclose pagination behavior, what an entry or diff looks like, or what happens when only one of the compare parameters is supplied.

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?

A single sentence with no filler, and the primary purpose (see history) is front-loaded before the secondary capability (compare). It is efficient, though for a tool with two distinct modes a slightly more explicit two-clause structure would have been clearer without adding length.

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?

This is a low-complexity, all-optional-parameter read tool with rich annotations and full schema coverage, so the description does not need to carry much. However, there is no output schema, and the description never indicates what a history listing or a comparison yields, nor how the two modes interact with the compare_* and starting_before parameters, leaving an agent to infer the return shape entirely.

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 100%, so all four parameters are already documented in the schema, including the compare_from/compare_to pairing and the starting_before pagination marker. The description's phrase 'two points in it' adds only a loose conceptual frame over the schema's explicit 'version number' language. With the schema doing the heavy lifting, the baseline of 3 applies.

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 names a specific resource and a specific verb: see the history of a site and compare two points in it. That is materially clearer than the bare name 'history' or the vaguer title 'See what changed', and it covers both the list mode and the diff mode. It stops short of naming a sibling (e.g. read_page or rollback) to contrast with, so it is clear but not sibling-differentiated.

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?

Usage is only implied: the phrase 'compare any two points' hints that supplying two version numbers switches the tool into diff mode, but there is no explicit statement of when to choose this over read_page, list_pages, or rollback. No exclusions or prerequisites are given, and it never says whether site can be omitted in multi-site connections (that lives only in the schema).

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

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

The annotations already declare readOnlyHint=true, idempotentHint=true, destructiveHint=false, and openWorldHint=false, so safety behavior is covered. The description adds useful return semantics by naming the fields returned. It does not mention pagination, truncation, ordering, or any limits on 'every page', which would matter for a list operation.

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?

It is a single, front-loaded sentence with no wasted wording. The scope and returned fields are stated immediately.

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 simple read-only list tool with complete schema descriptions and full safety annotations, the definition is nearly complete. It states the resource, scope, and returned fields. It could be stronger by clarifying whether results are paginated or complete, but the core invocation needs are covered.

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 100%, so both parameters are already well documented in the input schema. The description adds only the loose phrase 'in a site', which maps to the site parameter but does not add meaning beyond the schema. Baseline 3 is appropriate when the schema carries parameter semantics.

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 gives a specific verb and resource: 'List every page in a site.' It also states the returned fields (title, address, last changed), which clarifies the tool's scope. The phrase 'every page' implicitly distinguishes it from read_page, but no sibling is named explicitly.

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 says what the tool does but does not explain when to use it instead of read_page, find_references, or history. There is no guidance on when-not to use it, prerequisites, or how draft pages relate to the site inventory.

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

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

Beyond the mutation and destructive annotations, it discloses that changes are staged as a group, that a preview link is returned, and that nothing is kept at this step. Those details clarify the destructiveHint for delete operations, though durability and how the staged set is later applied remain unspecified.

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?

One compact sentence front-loads the staging action and ends with the preview-before-retention guarantee. The operation list and batching phrase earn their space, with no redundant explanation of schema fields.

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 three-parameter batch-mutation tool, the description plus fully documented schema supplies the core action, supported changes, staging behavior, and preview outcome; the returned link is especially useful because there is no output schema. The main remaining gap is the surrounding workflow—how this proposal relates to create_preview and how it is subsequently applied or published—so it is not fully complete.

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 100%, including operation types, body/path/title fields, draft_id provenance, and atomic operations, so the schema does the heavy lifting. The description reinforces the four operation kinds and batching but adds no parameter syntax or constraints beyond it, making the high-coverage baseline of 3 appropriate.

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 uses a specific verb and resource—"stage a set of changes"—and enumerates creates, edits, deletions, and moves. Its staged, preview-first outcome distinguishes it from siblings such as apply_edit and publish without requiring their schemas.

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 a clear use context: batch related page changes and inspect a preview before retaining them. It does not explicitly say when to choose apply_edit, create_preview, or publish, or what the next workflow step is, so it falls short of fully explicit alternative-based guidance.

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

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

Annotations flag destructiveHint=true and idempotentHint=false, but the description never explains what is destroyed (e.g., a previously published version is replaced) or that repeat calls produce different results. It also omits the rights_confirmed precondition, which the annotation profile alone cannot convey.

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?

It is a single front-loaded sentence with no filler, which is good structurally. However, the brevity is under-specification rather than earned concision given the destructive, multi-mode nature of the operation.

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 destructive, non-idempotent tool with four parameters and no output schema, the description should explain the two operating modes (publishing a draft preview vs. keeping a preview on a subdomain) and the consequences of publishing. None of that is present.

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 100%, so parameter meaning is already fully documented, including the draft_id-vs-site distinction and the subdomain case. The description adds no parameter semantics beyond the schema, which is the baseline expectation at full coverage.

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 gives a verb and resource ('put the current version of a site out for people to read'), which conveys roughly what publish means. But 'out for people to read' is a loose paraphrase that never says 'make publicly accessible' or 'replace the live version', and it does not distinguish this from siblings like create_preview or claim_draft.

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?

There is no when-to-use guidance, no prerequisites, and no mention of alternatives such as create_preview (which the schema references) or claim_draft. The agent must infer that create_preview must precede this call and that claim_draft is a different path.

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

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

Annotations already declare readOnlyHint, idempotentHint, destructiveHint=false and openWorldHint=false, so the safety profile is fully covered by structured data. The description adds only a weak hint that page-level 'top of page' content is included in the response. It does not explain what happens when a draft_id is supplied or how missing pages are reported.

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?

A single short sentence with the core action front-loaded and no redundancy. The trailing clause is the only weak spot because it is ambiguous rather than wasteful.

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?

The tool is simple and read-only, and the schema carries the parameter load well, but the description never addresses the draft-versus-live distinction that the draft_id parameter implies, nor what the returned page content includes. With no output schema, a slightly fuller description would help.

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 100%, and the schema descriptions are notably good, explaining that draft_id is the value from create_preview rather than the code in the preview link. The tool description adds no parameter meaning beyond that, so the baseline of 3 applies.

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 verb and resource are concrete ('Read one page of a site'), which distinguishes it from list_pages at a glance. But the qualifying clause 'including anything set at the top of it' is vague and does not say whether that means page-level settings, header content, or something else. No sibling is named or excluded.

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?

There is no statement of when to use this instead of list_pages, find_references, history, or create_preview/claim_draft for draft content. The only implicit cue is 'one page', which an agent must infer means a single-page fetch.

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

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

Annotations already declare destructiveHint=true, readOnlyHint=false and idempotentHint=false, so the safety profile is covered structurally. The description adds useful scope context (whole site by default versus a named subset of pages), which tells the agent what gets overwritten. It does not disclose whether the restored content goes live immediately, whether a new version is created, or whether the rollback itself can be undone, which is the key remaining behavioral question for a destructive tool.

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?

A single front-loaded sentence with no filler, and the site-versus-page scope appears before the temporal detail. It is efficient, though it is so terse that it carries almost no routing or safety information.

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?

Parameters are fully documented by the schema and the destructive profile by annotations, and there is no output schema to explain. What is missing for a destructive restore is whether the result is published or left as a draft, and how the agent should locate a valid version number, which the description leaves entirely 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 description coverage is 100%, so the baseline is 3 and the schema already explains site, pages and to_version. The description adds little beyond it and slightly understates the pages parameter, saying 'one page of it' when the schema accepts up to 500 page names.

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 sentence states a concrete verb ('put ... back') and resource ('a site, or one page of it') with a temporal qualifier ('an earlier point'), so an agent can tell it restores prior state rather than merely viewing it. It is clear on its own but never names the sibling that shows versions ('history') or contrasts itself with edit-oriented tools like 'apply_edit' or 'publish'.

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?

There is no when-to-use or when-not-to-use guidance: nothing says this is the tool for reverting unwanted changes versus editing forward, and no prerequisite is given (e.g. that a version number must first be obtained). The only hint at the workflow, 'from the history,' lives in the schema field for to_version, not in the description.

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

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

Annotations already flag destructiveHint=true and non-idempotency; the description adds the ownership constraint, the partial-update contract, and the non-obvious image gate — pictures are kept but not shown until the owner confirms them in the preview. It also notes that header links replace the current set, though the schema is what says this. It stops short of describing what a successful call returns or how failures surface.

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 is front-loaded and each sentence carries information, but the images_are_ours clause is long and largely restates the schema description, and the draft_id sentence is buried mid-paragraph. Dense run-on phrasing costs it some clarity.

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 12-parameter, all-optional mutation tool with no output schema, the description covers permissions, partial-update semantics, the preview path, image confirmation, and language scope. Gaps remain around return values and what happens to unmentioned settings, but the essential calling context is present.

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 already 100%, so the baseline is 3, but the description adds cross-parameter meaning: draft_id pairs with an unclaimed preview versus site for a kept site, images_are_ours is a permission assertion rather than a display toggle, and languages governs the site's own UI words like Previous/Next/Search. That goes beyond the per-field schema text.

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 ('Change how a site looks') and immediately enumerates the concrete facets: theme, main colour, typeface, header links, logo and icon. That enumeration distinguishes it cleanly from content-editing siblings like apply_edit, propose_edit, and publish without needing to name them.

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 real operational guidance: 'Only the site's owner can do this', 'Give only the things to change', and the rule that draft_id is used on an unclaimed preview while site is used for a kept site. It never explicitly contrasts itself with siblings such as apply_edit or propose_edit, so there is no stated when-not-to-use condition.

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

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

The annotations already declare readOnlyHint, idempotentHint, destructiveHint=false and openWorldHint=false, so the safety profile is fully covered without the description. The description does add value by disclosing what information comes back (including pending changes), which matters because there is no output schema, but the phrasing is metaphorical ('what it is on', 'waiting to be kept') rather than naming concrete fields.

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?

A single front-loaded sentence with four parallel clauses and no filler. It is appropriately sized for a read-only status check, though the figurative wording of the last two clauses costs some precision.

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 two-parameter read-only tool with full schema coverage and complete annotations, the description is close to sufficient: it covers what the tool reports. It leaves unclear how the site/draft_id combination changes behavior and how this differs from the many sibling read tools, which an agent would need to know to select correctly.

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 100%, and the parameter descriptions themselves explain the optional site and the draft_id semantics (the id from create_preview, not the code in the link). The tool description adds nothing about parameters, so it sits at the baseline where the schema carries the load.

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 names a specific verb ('check') and resource (a site) and enumerates what the check reports: published state, last publish time, current contents, and pending unpublished changes. It stops short of distinguishing this from siblings such as history or list_pages, which an agent might reach for when asking about a site's state.

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?

There is no explicit when-to-use guidance, no exclusions, and no mention of any sibling tool. The only routing hint is implicit: the schema's draft_id parameter suggests this also answers questions about previews, but the description never says when to prefer this over history or read_page.

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