케어그레이드 — 장기요양기관 평가 조회

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

Annotations already declare readOnly/idempotent/non-destructive with openWorldHint=false, so the safety bar is covered. The description adds genuinely new context: the data is NHIS-published and refreshed yearly, and missing values must not be guessed but routed to the url or contact page. It does not describe return layout or ordering, keeping it out of the top band.

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?

Short and largely front-loaded, with the comparison behavior leading. The bilingual duplication is justified for translation, though the English half drops the core purpose sentence and only carries the data-provenance/error guidance, which weakens the structure slightly.

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 no output schema, the description does address what comes back ('same fields side by side'), plus data provenance, update cadence, and missing-value handling, and annotations cover the safety profile. An agent has enough to call it correctly, with only minor gaps around result ordering.

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% with a single 'keys' parameter already documented as '비교할 기관 키 2~3개 / 2–3 keys', including min/max items. The description adds nothing about key format or eligibility, so the baseline 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?

"같은 필드를 나란히 돌려줍니다" (returns the same fields side by side) combined with the title makes the comparison verb and resource explicit, and it is distinguishable from single-institution siblings like get_institution. It stops short of naming the sibling it replaces, so it is clear but not maximally 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?

The description gives a real conditional instruction — do not let users compare grades across different benefit types — but this is post-processing/presentation advice, not when-to-call guidance. It never states when to prefer this tool over get_institution or search_institutions, so usage is only implied.

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 cover the safety profile (readOnly, idempotent, non-destructive, closed-world), so the bar is lower, and the description still adds meaningful traits: data provenance (NHIS-published), a yearly refresh cadence implying staleness, and the rule that suspended institutions are reported as absent rather than erroring. It omits nothing critical but does not discuss pagination or error responses.

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 return fields are front-loaded in the first sentence, followed by data-freshness and missing-value rules, with no filler. The bilingual duplication is purposeful for the audience, though it roughly doubles the 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?

With no output schema, the description carries the return-value burden and does so by listing the fields returned. Combined with the missing-value and freshness disclosures, an agent has enough to call it correctly, with only the hand-off to search_institutions left unstated.

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 single 'key' parameter already documents its format, pattern and source path (/c/<키>). The description only restates that lookup is by 기관 키, adding no syntax or format detail beyond the schema — the baseline 3 for schema-documented parameters.

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 ('기관 키로 한 곳의 ... 돌려줍니다') and enumerates the returned fields (등급·평가연도·주소·인력·정원·등급 이력), so an agent knows exactly what it gets. The '한 곳' (single record) framing implicitly separates it from search_institutions/compare_institutions, 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?

Gives real operational guidance — suspended institutions answer as absent, and missing values must not be guessed but routed to the url or contact page. However, it never states when to choose this tool over its close sibling search_institutions (e.g. when the key is unknown), leaving the routing to inference.

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, so safety is covered. The description adds value beyond that: values are NHIS-published, updated only once per year (freshness), and it prescribes behavior for gaps (don't fabricate, redirect to url/contact). That is genuine behavioral context.

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?

Content is front-loaded with the returned fields, followed by freshness and the no-guessing rule. The bilingual duplication roughly doubles the length, but each sentence carries distinct information with no filler.

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 no output schema, the description appropriately compensates by enumerating the returned fields, plus freshness and missing-value handling. Complete enough for an agent to call a zero-parameter stats tool correctly, though a hint about the value scale or format would round it out.

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?

The tool takes zero parameters, so the baseline is 4. The description lists the data categories surfaced, which is more useful than anything a parameter schema could add here.

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?

It clearly states the tool returns site-wide aggregate statistics and enumerates the specific returned items (data baseline date, national institution count, grade distribution, counts by benefit type, source/disclaimer). This distinguishes it from institution-level siblings like get_institution and search_institutions, though it does not name an alternative 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?

Usage is implied by the aggregate nature of the resource, and it gives a handling rule ('never guess missing values; point to the url or contact page'), but it never states when to prefer this over siblings like compare_institutions or list_regions. No explicit when/when-not routing.

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 is covered. The description adds genuinely useful non-annotation context: the values are NHIS-published yearly-refreshed evaluation results (data provenance and staleness) and an explicit instruction not to fabricate missing values. It does not describe pagination or result size.

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 content is front-loaded and each sentence carries information, but the entire text is duplicated verbatim in Korean and English, roughly doubling the length with zero additional semantic payload. It is adequate rather than tight.

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?

There is no output schema, and the description compensates reasonably by describing what the result contains (institution counts by region, benefit-type counts, embedded region page URL). Combined with the yearly-refresh provenance note, an agent has enough to call and interpret it, though result shape details remain thin.

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, including that omitting sido yields province totals. The description's restatement of the gugun-plus-sido behavior adds no syntax or format detail beyond the schema, 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?

States a specific resource and measure: institution counts by province/district, plus benefit-type breakdown when a district is supplied, and mentions the region page URL is included in the result. It is distinguishable from get_stats and search_institutions, though it never names or contrasts with those siblings.

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 conditional behavior ('if you pass gugun you get benefit-type counts' and 'never guess missing values, point to the url or contact'), but it does not say when to prefer this over get_stats or search_institutions, nor any prerequisites or exclusions. Usage is implied rather than stated.

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 cover safety (readOnly, idempotent, openWorld=false), and the description adds real value beyond them: the exact return shape (title/summary/date/url, body deferred to url), the data provenance (NHIS-published, yearly refresh), and an anti-hallucination rule for missing values.

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 return-shape statements are front-loaded and useful, but the middle section about NHIS evaluation values and yearly updates reads like boilerplate borrowed from a different data-lookup tool and does not obviously belong to an article search.

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?

No output schema exists, so the description correctly takes on the burden of describing the returned fields and the url-deferral behavior, which makes it largely callable. However, it omits the undocumented limit/category parameters and mixes in provenance text that poorly fits an article-search tool.

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 only 33% (just 'q'), and the description does not compensate: it never mentions the 'limit' cap of 10 or the 'blog'/'news' category filter. Only the keyword scope ('제목·요약에서 찾는 말') indirectly maps to q.

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?

Names a specific verb and resource ('안내 글·뉴스 찾기 / Search articles') and explicitly states what is returned (title, summary, publish date, url) versus what is deferred to the url. It is clearly distinguishable from get_institution/compare_institutions, though it never explicitly contrasts itself with the closest sibling search_institutions.

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 choose this tool over search_institutions or search_institutions-style lookups; the only usage-like instruction is about handling missing values ('point to the url or the contact page'), which is a fallback rule, not a selection rule.

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 readOnly/idempotent/non-destructive/openWorld=false, so safety is covered; the description adds genuinely new context: values are NHIS-published evaluation results refreshed once a year, and missing values must not be guessed but referred to the page url or contact page. It does not mention result caps or paging behavior despite the limit parameter's max of 20.

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 action is front-loaded and both language versions are tightly written with no filler. Slight inefficiency in that the English half drops the filter enumeration and return fields present in the Korean half, so the bilingual pair is not fully parallel.

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 no output schema, the description correctly fills that gap by naming the return fields (grade, evaluation year, address, page url), and it adds provenance plus an anti-hallucination rule for missing data. It leaves result-set size/paging behavior around the limit parameter unstated, which is the main remaining gap for a search tool.

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 83%, so the schema already documents q, svc, sido, grade and gugun; the description largely restates those same filter axes and the 'code or name' / 'name or region substring' semantics already present in the schema. The limit parameter is not addressed anywhere in the description, so it adds little beyond the documented schema.

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 concrete verb+resource (find/filter long-term care institutions) plus the exact filter axes (name/region, province, district, service type, grade) and what comes back (grade, evaluation year, address, page url). The multi-result search/filter scope clearly separates it from get_institution and compare_institutions 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?

Usage is implied by the filter enumeration ('find by name/region or filter by province/district/service/grade') and by the directive on missing values, but the description never names an alternative tool or states when a caller should prefer compare_institutions, get_institution, or list_regions. Adequate but leaves routing to inference.

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