Korean accounting standards (K-IFRS, K-GAAP, audit, ICFR) and official Q&A, cited by paragraph.
Scored across 11 tools
Most tools have clearly distinct purposes (search vs list_* vs get_topic vs verify_quote vs find_evidence), and the descriptions helpfully cross-reference each other. However, the generic `fetch` (by id) substantially overlaps with the more specific `get_qna` (by slug), `get_standard`, and `get_paragraphs`, so an agent could reasonably hesitate over which retrieval tool to pick.
The set mostly follows a predictable verb_noun convention with a clear `get_*` family (get_paragraphs, get_qna, get_related, get_standard, get_topic) and a `list_*` family (list_standards, list_audit_cases), plus find_evidence/verify_quote. The only minor deviations are the bare single-word verbs `fetch` and `search`, which are still verb-style and readable.
Eleven tools is well within the ideal 3-15 range and each tool maps to a distinct facet of the accounting-wiki domain (discovery, retrieval, aggregation, verification). No tool appears redundant or gratuitous at the count level.
The surface covers the full research lifecycle for this domain: discovery (search, list_standards, list_audit_cases, get_topic), retrieval by multiple keys (fetch, get_paragraphs, get_qna, get_standard), evidence aggregation and linkage (find_evidence, get_related), and quote verification (verify_quote). No obvious dead ends or missing operations for an accounting reference wiki.
회계위키에서 search 등이 반환한 id(para:… / qna:… / std:… / article:…)로 해당 문단·질의회신·기준서 목차·실무해설의 전문 마크다운을 조회. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | search 결과의 id (예: 'para:ifrs/1116/main/p22') |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnly/idempotent/non-destructive/non-open-world, so the safety profile is covered. The description adds genuinely useful non-schema context: results are for lookup purposes only, copyright vests with the issuing body, and the source URL must be cited — constraints an agent would otherwise not know.
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?
Three sentences, front-loaded with purpose then retrieval scope, with licensing/citation boilerplate trailing. The copyright and citation sentences are slightly peripheral to invocation but are short and legally purposeful, so there is little waste.
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 states the return format (full markdown of the document) and the id input contract, and adds citation obligations. Given a single-parameter read tool with full annotation coverage, this is adequately 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 coverage is 100% and the id pattern is already exemplified in the schema, so baseline is 3. The description goes further by enumerating the four id namespaces (para:/qna:/std:/article:) and mapping each to a content type, which clarifies what argument shapes are valid.
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 (조회/retrieve) and resource (문단·질의회신·기준서 목차·실무해설 full markdown), and identifies the key (search-returned id). It is clear what comes back, though it never differentiates itself from the typed siblings get_paragraphs/get_qna/get_standard that appear to cover the same id prefixes.
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 tells the agent where the argument comes from (search 등이 반환한 id), which is a usable precondition. However, it gives no guidance on when to pick this generic fetch over get_paragraphs/get_qna/get_standard/get_topic for a given id prefix, leaving the main routing decision to inference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계 질문 하나에 대해 회계위키의 근거를 한 번에 묶어 반환: 핵심 기준서 문단(앞뒤 문단 요약 포함), 관련 질의회신·감리지적사례(요지와 인용한 기준서 문단), 실무해설. 답변을 쓰기 전 근거 수집용. 질문은 문장 그대로 넣어도 되며, 결과에 '근거 약함' 주의가 붙으면 기준서 용어로 바꿔 다시 호출할 것. 각 항목의 id 로 fetch·get_paragraphs 를 부르면 전문을 얻는다. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| std | No | 기준서 번호로 한정 (예: '1116') | |
| domain | No | 도메인 한정 — 생략 시 전체 | |
| question | Yes | 회계 질문 (예: '임대인이 임대료를 깎아주면 임차인은 어떻게 회계처리하나') | |
| max_chars | No | 응답 목표 길이(자), 기본 10000 |
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), and the description adds materially beyond them: the shape of the returned bundle, a weak-evidence quality signal with a retry instruction, and licensing/citation obligations. These are behavioral traits an agent cannot infer from structured 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?
Front-loaded with the core purpose and content inventory, then usage, retry guidance, and citation rules in a logical order. It is a dense single block rather than a list, and the licensing sentence is long, but every sentence carries actionable 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?
There is no output schema, so the description must describe the return, and it does so concretely (standard paragraphs plus neighbors, Q&A/audit cases with cited paragraphs, practical commentary, source URLs, per-item ids). Combined with the retry and citation guidance, an agent has everything needed to call and follow up 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%, so std, domain, max_chars and question are already documented with examples and ranges. The description only adds that a natural-language sentence can be passed as-is, which slightly enriches the question parameter but does not compensate for anything since nothing is missing. 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?
States a specific verb and resource ('회계위키의 근거를 한 번에 묶어 반환') and enumerates exactly what the bundle contains: standard paragraphs with surrounding context, related Q&A/audit cases, and practical commentary. It is clearly differentiated from siblings like fetch/get_paragraphs, which it explicitly names as the full-text follow-up.
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 an explicit when-to-use ('답변을 쓰기 전 근거 수집용'), tells the agent that a verbatim question is acceptable, and prescribes a concrete recovery path when the result carries the '근거 약함' warning (re-query with standard terminology). It also routes to the follow-up tools (fetch, get_paragraphs) via item ids.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계위키에서 기준서의 특정 문단들을 번호로 조회 (예: std='1116', paras=['22','B34']). 적용지침(B)·결론도출근거(BC) 문단도 토큰으로 자동 해석. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| std | Yes | 기준서 번호 (예: '1116', kgaap 은 '13') | |
| paras | Yes | 문단 번호 목록 (예: ['22','한54.2','B34','BC78']) | |
| domain | No | 기본값 ifrs |
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. The description adds genuinely useful behavior: B (적용지침) and BC (결론도출근거) paragraph tokens are auto-resolved, and results are query-only with an attribution obligation to cite the source URL — neither is derivable from the schema.
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?
Front-loaded with purpose and a usage example in the first sentence, followed by the token-resolution rule and the citation requirement. Four sentences, all relevant, though the copyright/attribution sentence is somewhat boilerplate relative to the core lookup instruction.
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 lookup with full schema coverage and no output schema, the description covers purpose, example format, token behavior, and citation duty. Only gaps are error/empty-result handling and explicit sibling routing, which are minor here.
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. The description goes slightly beyond by showing the std/paras value shapes and, importantly, explaining that paras entries with B/BC prefixes are auto-interpreted — semantics the schema's generic '문단 번호 목록' does not convey.
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 (조회/retrieve) and resource (기준서의 특정 문단) and narrows scope to lookup by paragraph number, with a concrete example (std='1116', paras=['22','B34']). It implicitly differs from get_standard (whole standard) and search, but never names or contrasts those siblings 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 example and the 'specific paragraphs by number' framing, so an agent can infer this is for targeted paragraph retrieval. However, it never states when to prefer this over get_standard or search, nor what to do if a paragraph number is unknown or invalid.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계위키의 질의회신·감리지적사례 전문 조회 (slug 는 search 결과 또는 페이지 URL 마지막 경로). 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| slug | Yes | 질의회신 슬러그 (예: 'SSI-38683') | |
| domain | No | 기본값 ifrs |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover readOnly, idempotent and non-destructive, so the safety profile is handled. The description adds non-obvious context beyond them: the content is full text, it is licensed by the issuing body, and citations must include the source URL. It still omits return shape/pagination, hence not a 5.
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?
Front-loaded with the lookup purpose in the first clause, followed by the slug-source hint and a compliance sentence. Dense but no filler; the citation/copyright sentence is boilerplate yet defensible for a legal-content tool.
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 still tells the agent it receives full text for a given slug, plus the citation obligation, which is the operationally important part. Nothing else is needed to call a two-parameter, fully documented read tool 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 coverage is 100%, so baseline is 3, and the description goes one step further by explaining the provenance of the required slug (search result or URL path segment), which the schema only illustrates with an example. The optional domain enum is left to the 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?
Names a specific verb (전문 조회) and a specific resource type (질의회신·감리지적사례), which is enough to distinguish it from list_audit_cases (a listing) and search (discovery). It does not explicitly contrast with siblings like get_standard or get_topic, but an agent can tell what content this returns.
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 documents where the required slug comes from ('search 결과 또는 페이지 URL 마지막 경로'), which is a genuine prerequisite and implies calling search first. However, there is no explicit statement of when to prefer this tool over get_topic/get_standard or any when-not condition.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계위키에서 기준서의 목차(섹션·문단 범위)와 메타데이터 조회. section 지정 시 해당 섹션 본문 전체를 반환. 전문 마크다운은 응답의 .md URL 로도 제공. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| std | Yes | 기준서 번호 (예: '1116') | |
| domain | No | 기본값 ifrs | |
| section | No | 섹션 제목 (목차에 나온 제목, 부분 일치 허용) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare the safe read-only, idempotent profile, so the bar is lower, yet the description adds genuine context beyond them: specifying section returns the entire section body, full markdown is also reachable via a .md URL in the response, and results are for lookup only with a citation obligation to the issuing body. It does not mention pagination or size limits, which keeps it short of a 5.
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?
Four compact sentences, front-loaded with the core retrieve behavior and only then the section-body effect, markdown URL, and copyright/citation constraint. Nothing is wasted, though the licensing sentences are boilerplate that could be trimmed.
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 burden of describing returns and does so: TOC with section/paragraph ranges, metadata, full section body on demand, and a .md URL for the whole document. Partitioning versus sibling tools like get_paragraphs is the one thing 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 100%, so the baseline is 3 and the schema already documents std, the domain enum, and section's partial-match behavior. The description nevertheless adds meaning the schema lacks — that supplying section changes the response from a TOC to the full section body — which justifies going above 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?
The description states a specific verb+resource: retrieving a standard's table of contents (sections and paragraph ranges) plus metadata from the accounting wiki, and notes it returns the full section body when a section is named. That is enough to distinguish it from list_standards (listing) and get_paragraphs (paragraph-level retrieval), though it never names those siblings 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 — reach for this when you need a standard's TOC/metadata or a whole section body — but there is no explicit when-to-use/when-not guidance and no routing to alternatives such as get_paragraphs for granular paragraph lookups. A reader must infer the boundary from the purpose statement alone.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계위키의 회계 주제(예: '자산손상', '리스', '수익')의 정의·관련어, 핵심 기준서 문단 범위, 대표 질의회신·감리지적사례·실무해설을 한 번에 반환. 주제명·관련어·slug 로 조회하며, topic 을 생략하면 전체 토픽 목록. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | 질의회신·실무해설 종류별 최대 건수, 기본 8 | |
| topic | No | 주제명·관련어·slug (예: '자산손상', '손상차손', 'asset-impairment') |
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. The description still adds non-obvious context: the omitting-topic fallback to a full list, and the licensing/citation constraint that results are for query purposes only and the source URL must be cited. It does not describe pagination or the shape of the aggregated payload.
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?
Three sentences, front-loaded with what gets returned, then lookup keys, then the legal/citation note. Dense but each sentence carries distinct information; the copyright sentence is slightly tangential to tool selection but is a genuine usage constraint.
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 usefully enumerates the returned content types and the optional-topic fallback, which is what an agent needs to know before calling. Minor gaps remain around result size/pagination and how the aggregated sections are structured, but the core is 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 coverage is 100%, so both parameters are already documented, making 3 the baseline. The description restates the topic lookup keys (name/related term/slug) that the schema already lists and says nothing about the limit parameter's per-type cap or default beyond what the schema provides.
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 (회계 주제 / accounting topics) and enumerates exactly what is returned: definition and related terms, key standard paragraph ranges, representative Q&A, audit-inspection cases, and practical commentary. An agent can tell it is a topic-level aggregator, though it never explicitly contrasts itself with siblings like get_standard, get_qna, or list_audit_cases, which overlap in content.
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: query by topic name/related term/slug, and omitting topic yields the full topic list. There is no explicit guidance on when to prefer this aggregator over the narrower sibling tools (get_qna, get_standard, get_paragraphs) or any stated exclusions, leaving the routing decision to inference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계위키에 수록된 금융감독원·한국공인회계사회 감리지적사례를 계정·위반유형·기관·연도·키워드로 걸러 최근 순 목록을 반환. 조건 없이 호출하면 계정·위반유형 분류와 건수를 반환. 목록의 id 를 fetch 에 넘기면 사례 전문. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | 기본 20 | |
| query | No | 제목·본문 키워드 (부분 일치) | |
| source | No | 기관 — '금감원' 또는 '한공회' | |
| account | No | 계정·쟁점 (부분 일치, 예: '재고자산', '매출') | |
| year_to | No | 종료 연도 | |
| violation | No | 위반유형 (부분 일치, 예: '과대계상', '주석 미기재') | |
| year_from | No | 시작 연도 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnly/idempotent/non-destructive, so the safety profile is covered. The description adds real behavioral context beyond that: results are ordered most-recent-first, the unfiltered call returns aggregates rather than records, and there is a usage/copyright restriction requiring source-URL citation. Return format and pagination limits are left implicit.
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?
Front-loaded with the core purpose, then the no-filter case, then the fetch hand-off, then the legal note. Sentences earn their place, though the copyright/citation sentence is somewhat tangential to invoking the tool.
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 usefully explains what comes back (a recent-ordered list, or classification+counts when unfiltered) and how to get full text. It omits return shape details like pagination or the fields of each list item, which keeps it just short of 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%, so every parameter including the default limit of 20 is already documented in the schema. The description restates the filter dimensions but adds no syntax, matching rules, or format guidance beyond what the schema provides.
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 verb (반환/목록) plus the exact resource (회계위키 수록 금감원·한공회 감리지적사례) and the filterable dimensions (계정·위반유형·기관·연도·키워드). It also distinguishes behavior from a sibling by stating that passing a list id to fetch yields the full case text, so an agent can tell list vs. detail apart.
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 clearly explains the no-filter branch (returns classification and counts) and routes the agent to fetch for full text via the id. It stops short of explicitly saying when NOT to use this tool or naming other alternatives among the many sibling search/list tools.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계위키에 수록된 도메인의 기준서 전체 목록 (번호·제목·문단수·URL). 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| domain | No | 기본값 ifrs |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare the safety profile (readOnly, idempotent, non-destructive, closed-world), so the description need not restate it. It does add genuine non-schema context: results are lookup-only, copyright rests with the issuing body, and the source URL must be cited in responses – a real behavioral obligation. It omits any note on result size or pagination for a full-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?
Three sentences, purpose front-loaded, and the returned-field list is compact. The copyright/citation sentences are somewhat verbose but carry legitimate licensing context rather than 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 usefully enumerates the returned fields (number, title, paragraph count, URL) and the domain scope, which is enough for an agent to call and consume it. The only gap is the absence of any statement about result volume or pagination for a 'full list'.
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 single 'domain' enum already documents its values and default (ifrs). The description only alludes to domain scoping via '도메인의' and adds no format or default guidance 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 verb+resource ('기준서 전체 목록' – full list of standards) and enumerates the returned fields (번호·제목·문단수·URL), which distinguishes it from the singular get_standard sibling. It does not, however, explicitly name or contrast itself with the other retrieval tools (search, get_topic).
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: as a '전체 목록' listing tool, an agent can infer it is for browsing/discovery before drilling into a specific standard. There is no explicit when-to-use statement, no mention of alternatives like search or get_standard, and no exclusion conditions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
회계위키 전체(기준서 문단·질의회신·감리지적사례·실무해설) 검색. 키워드와 뜻(의미)을 함께 보므로 '리스료를 깎아주면 어떻게 처리하나'처럼 문장으로 물어도 된다. 기준서 용어·문서번호는 정확히 일치하는 결과가 위로 온다. 실무 표현으로 찾은 결과가 약하면 기준서 용어로 바꿔 한 번 더 검색할 것(예: '계약 따낸 대가로 준 수수료' → '계약체결 증분원가', '앱마켓 수수료 빼고 매출' → '본인 대 대리인'). 질문에 대한 근거를 한 번에 모으려면 find_evidence 가 낫다. 결과의 id 를 fetch 에 넘기면 전문을 얻는다. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| std | No | 기준서 번호로 한정 (예: '1116') | |
| kind | No | standard: 기준서 문단만, qna: 질의회신·감리사례만, article: 실무해설만 | |
| limit | No | 기본 10 | |
| query | Yes | 검색어 (예: '리스 변경 회계처리') | |
| domain | No | 도메인 필터 — 생략 시 전체 도메인 검색 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnly/idempotent/non-destructive, so the description adds ranking behavior (exact standard terms and document numbers rank higher), semantic matching behavior, and copyright/citation obligations. Missing detail on result count defaults beyond schema, but schema covers limit; strong added value.
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?
Front-loads scope then matching behavior then workflow guidance. Mostly efficient, though the copyright/citation sentence is somewhat tangential and could be condensed; overall well-structured with no true waste.
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 read-only search tool with no output schema, the description covers scope, matching semantics, ranking, alternatives, follow-up via fetch, and citation obligations. Return value shape (fields like id) is only implied but sufficient for correct invocation.
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 enum descriptions already explain kind and domain. The description adds no parameter-level detail beyond what the schema provides; baseline 3 is 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?
States a specific verb (검색) and scope (회계위키 전체 with enumerated content types: 기준서 문단·질의회신·감리지적사례·실무해설). Distinguishes from siblings by naming find_evidence for evidence gathering and fetch for full text.
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?
Explicitly says when to use it (semantic/phrase queries, exact term matching ranks higher), when to use an alternative (find_evidence for collecting evidence for a question), and gives a concrete retry strategy with examples of converting practice phrasing to standard terms.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
기준서 문구 인용이 회계위키 원문과 일치하는지 대조. 공백·문장부호 차이는 무시하고 일치(exact)·표현 차이(near)·부분 일치(partial)·불일치(mismatch)로 판정하며, 일치하지 않으면 가장 가까운 원문 문장과 그 문단을 반환. para 를 생략하면 기준서 전체에서 인용문이 있는 문단을 찾는다. 결과는 이용자 질의에 대한 조회 목적으로만 제공되며 원문 저작권은 제정·발행기관에 있다. 인용 시 응답의 출처 URL 을 표기할 것.
| Name | Required | Description | Default |
|---|---|---|---|
| std | Yes | 기준서 번호 (예: '1116', kgaap 은 '13') | |
| para | No | 문단 번호 (예: '22', 'B34') — 생략 시 기준서 전체 | |
| quote | Yes | 대조할 인용 문구 (5자 이상) | |
| domain | No | 기본값 ifrs |
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), and the description adds real behavioral detail beyond them: whitespace/punctuation are normalized before comparison, the verdict taxonomy, and fallback behavior returning the closest sentence and paragraph. The copyright/attribution sentence adds an output-usage constraint. Doesn't describe pagination/rate behavior, but the bar is lower given 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?
Front-loads the core comparison behavior, then layers in matching semantics and the para-omission rule efficiently. The closing copyright/attribution sentence is slightly tangential but earns its place as an agent-facing output constraint.
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, so the description carries the burden of explaining returns — and it does, listing verdict labels and the returned nearest sentence plus paragraph. Complete enough to call the tool correctly, though it omits any statement on scoring/output shape beyond the labels.
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%, so all four parameters are already documented, and the description's note that omitting para searches the whole standard merely restates the schema's para description. No additional syntax or examples are supplied beyond what the schema provides.
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 (대조/compare) and two concrete resources (기준서 문구 인용 vs 회계위키 원문), and enumerates the judgement outcomes (exact/near/partial/mismatch). This clearly separates it from the retrieval siblings (get_paragraphs, get_standard, search).
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 useful context — the para-omitted case searches the whole standard, and mismatches still return the nearest sentence — but it never says when to prefer verify_quote over siblings like get_paragraphs or search. Usage is implied rather than routed.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
find_evidenceget_relatedget_topiclist_audit_casessearch2 fields changedInput schema / properties / kind / descriptionPrevious value: -"standard: 기준서 문단만, qna: 질의회신·사례만"New value: +"standard: 기준서 문단만, qna: 질의회신·감리사례만, article: 실무해설만"Input schema / properties / kind / enumPrevious value: -[
- "standard",
- "qna"
-]New value: +[
+ "standard",
+ "qna",
+ "article"
+]verify_quotefetchget_paragraphsget_qnaget_standardlist_standardssearchKorean accounting standards, audit, tax and legal sources and rulings, with citation checks
Full-text search over K-IFRS/K-GAAP standards, Q&A and KASB/KICPA material for Korean accountants
Full-text search over FSS/FSC accounting supervision documents for Korean accounting professionals
Korean policy & finance facts with primary-source citations (laws, gov notices) for AI agents
Add one secure layer between your agents and this server.