Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mfds-mcp Search for current guidelines on food additives"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
식약처(MFDS)의 법령정보(게시판 m_203)와 공무원 지침서/민원인 안내서(게시판 m_1059)를
크롤링·캐시하고, RA(인허가) 업무에서 "현행/최신 추정본을 근거와 함께" 검색·반환하는 MCP 서버.
Context7가 최신 라이브러리 문서를 정확히 뽑아주듯, 식약처 자료에서 폐지·구버전을 걸러 현행 추정본을 제공한다. 단, 현행임을 단정하지 않고 상태·근거·신뢰도를 함께 반환한다 (잘못된 "현행" 단정은 검색 실패보다 위험하다).
두 게시판 크롤링 (공개 API 없음 → HTML 크롤 + 첨부 파싱)
m_1059 지침서: view.do?seq=N 상세 + HWP/HWPX/PDF 첨부가 실제 내용 → 핵심 큐레이션 영역
m_203 법령: 각 항목이 외부 law.go.kr 로 연결되는 개정공포 피드
첨부 전문 텍스트 추출 → SQLite FTS5(trigram) 한국어 전문검색
버전 그룹핑 + 상태 판정: current / superseded / future / repealed / unknown + 근거 + 신뢰도
Related MCP server: LexLink
m_203은 개정공포 피드이지 현행 법령 본문 DB가 아니다. 따라서 본 서버의 법령 검색은
**"개정공포 검색"**이며, 각 결과의 법령 현행 여부는 항상 law_currency = "unknown"으로 반환하고
law.go.kr 원문 링크로 확인하도록 안내한다. "옛 시행령 필터링(현행 본문 제공)"의 완성은
Phase 2: law.go.kr 현행법령 공개 API 연동(무료 OC키)에서 이뤄진다.
uv sync # 또는: pip install -r requirements.txt레거시 .hwp 전문추출 품질을 높이려면 (선택, 없어도 olefile+zlib 네이티브 경로로 동작):
uv sync --extra hwp로컬(Claude Desktop, stdio):
uv run mcp_server.pyClaude Desktop/Code 설정은 mcp.json.example 참고. DB 경로는 MFDS_DB_PATH(선택, 기본 mfds_cache.db).
원격(Claude.ai 웹 + Desktop, 클라우드): → DEPLOY.md 참고.
Prefect Horizon(구 FastMCP Cloud, 콘솔 horizon.prefect.io)에 GitHub 레포를 연결하면
https://<이름>.fastmcp.app/mcp URL이 생기고, Claude 설정 → 커넥터에 등록하면 웹/데스크톱
어디서나 쓸 수 있다. 엔트리포인트는 mcp_server.py:mcp. 임시 파일시스템 대응으로 시드 DB(mfds_seed.db) + 서버 내 백그라운드
자동 크롤을 사용한다.
로컬 HTTP 테스트:
MFDS_TRANSPORT=http MFDS_HTTP_PORT=8000 uv run mcp_server.py # http://localhost:8000/mcp검색 도구는 읽기 전용이다. 데이터 적재/갱신은 별도 운영 명령으로 실행한다:
uv run admin.py refresh m_1059 --max-pages 5
uv run admin.py refresh m_203 --max-pages 5
uv run admin.py reindex # FTS 재색인
uv run admin.py status # 진단(문서수/파싱실패율/마지막 크롤)(MCP 도구 mfds_refresh 로도 가능하나 느리고 쓰기 작업이므로 권장하지 않음.)
도구 | 용도 |
| 지침서(m_1059) 전문검색, 현행본 우선 + 상태/근거 |
| 법령(m_203) 개정공포 검색 + law.go.kr 링크 |
| 단일 문서 조회 (meta/full/페이지 청크) |
| 주제→현행 후보 + 근거·신뢰도 + 이력 |
| 그룹 전체 판본·상태 감사(audit) |
| 최근 제·개정 모니터링 |
| 키워드 주변 발췌 + 출처(첨부·페이지) |
| 진단(크롤 시각·문서수·파싱실패율) |
| (쓰기) 크롤·갱신 — 운영용 |
데이터 출처: 식품의약품안전처(www.mfds.go.kr). 크롤링은 보수적 호출 빈도로 수행하며, 공공누리 등 출처 이용조건을 준수한다. 본 도구는 참조 보조용이며, 규제 판단의 최종 확인은 반드시 원문(첨부 원본 / law.go.kr)으로 해야 한다.
mfds_client.py HTTP 크롤 (목록/상세/첨부 다운로드, charset·UA·백오프·DOM검증)
parsers.py HWP/HWPX/PDF → 텍스트 (+페이지경계, 품질점수)
versioning.py 그룹핑·상태판정·근거/신뢰도
db.py SQLite upsert/FTS/조회 + override 반영
services.py 중수준: 크롤+캐시, 검색, 최신버전 해석
mcp_server.py FastMCP(v2/v3) 도구 등록 + 백그라운드 자동크롤 (엔트리포인트 mcp_server.py:mcp)
admin.py 운영 CLI (refresh/reindex/status/override)
selftest.py 결정적 단위테스트 + 라이브 스모크
schema.sql SQLite 스키마 (FTS5 trigram)
mfds_seed.db 시드 데이터 (배포 직후 즉시 검색용; 런타임 캐시는 mfds_cache.db)
DEPLOY.md 클라우드/원격 배포 가이드긴 지침서 본문에서 키워드 주변 섹션만 발췌(출처 첨부명 포함).
| Name | Required | Description | Default |
|---|---|---|---|
| board | Yes | ||
| seq | Yes | ||
| keyword | Yes | 찾을 키워드. 예) "독성시험", "가용화제" | |
| window | No | 발췌 길이(문자). |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true, so the description adds context like including source attachment name. However, it does not disclose details about behavior like truncation or error handling.
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?
Single sentence with key information (action, resource, source inclusion) is efficient and front-loaded, though extra parameter hints could help.
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?
Lacks explanation for required parameters board and seq, which are essential for identifying the document. With output schema present, the description should clarify these identifiers.
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 50% (keyword and window have descriptions), but the description does not add meaning for the undocumented required parameters board and seq. It only implies keyword usage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Description clearly states specific verb 'extract' and resource 'sections around keywords from long guideline text', and distinguishes from siblings like mfds_get_document (full document) and mfds_search_guidelines (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?
The description implies usage for extracting keyword-related sections, but does not explicitly state when to use this tool versus alternatives or provide any exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
단일 문서 조회. 컨텍스트 절약을 위해 part='meta'(기본) 또는 'full'.
| Name | Required | Description | Default |
|---|---|---|---|
| board | Yes | 'm_1059' | 'm_203' | |
| seq | Yes | 문서 seq | |
| part | No | 'meta'(메타+첨부목록) | 'full'(본문 텍스트 포함) | meta |
| max_chars | No | full 일 때 본문 최대 길이. |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true, so the description does not need to reiterate. The description adds context about the part parameter's role in saving context, but beyond that, no behavioral details (e.g., pagination, error handling) are provided.
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, well-structured sentence that immediately states the main purpose and follows with a key usage tip. Every word serves a purpose, and there is no redundancy.
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?
Given the presence of a full output schema (not shown) and complete parameter descriptions, the description is sufficient for a straightforward tool. It clearly communicates the core action and a critical parameter nuance. It could be slightly improved by clarifying what 'meta' includes, but it is not necessary.
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 input schema covers all 4 parameters with descriptions. The description adds semantic value by explaining the rationale behind the part parameter default ('컨텍스트 절약을 위해'), which helps an agent understand when to use 'meta' vs 'full'.
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 clearly states the action (단일 문서 조회, i.e., view a single document) and the resource (문서). It is distinct from sibling tools like mfds_list_recent or mfds_extract_section, which serve different purposes.
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 implies usage for retrieving a single document and suggests using part='meta' to save context, but it does not explicitly state when to choose this tool over siblings or provide exclusion criteria.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
주제/제목으로 현행 지침서 후보 1건을 근거·신뢰도와 함께 해석 + 판본 이력.
반환: current(현행 추정본), current_confidence, alternatives, version_history. ⚠ current 는 자동 추정. confidence 낮거나 status=unknown 이면 원문 확인 필수.
| Name | Required | Description | Default |
|---|---|---|---|
| title_or_topic | Yes | ||
| domain | No |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations declare readOnlyHint=true, so the description does not contradict. The description adds valuable context: the tool auto-estimates the current version, confidence levels, and warns about low confidence or unknown status requiring manual verification.
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?
Description is very concise: action stated first, return fields listed, then a warning. No wasted words. Well-structured for easy scanning.
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?
Output schema exists, so return values are documented. The description covers primary behavior, confidence estimation, version history, and verification warning. It lacks mention of domain parameter handling and error conditions, but overall 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 0%, so description must compensate. It provides context for 'title_or_topic' (주제/제목) but does not explain the optional 'domain' parameter. Missing parameter semantics reduce clarity for the agent.
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 clearly states the tool returns the current guideline candidate with evidence and confidence, plus version history, based on a subject/title. It distinguishes this from siblings like mfds_version_history (which returns full history) and mfds_search_guidelines.
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 implies usage when needing the latest version with confidence assessment, but does not explicitly state when to avoid this tool or mention alternatives like mfds_get_document or mfds_search_guidelines.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
최근 제·개정 모니터링. board별 최근 문서를 게시일 내림차순으로.
| Name | Required | Description | Default |
|---|---|---|---|
| board | Yes | 'm_1059' | 'm_203' | |
| days | No | 최근 N일(0이면 무제한) | |
| domain | No | 도메인 필터(선택) | |
| limit | No |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The annotation indicates readOnlyHint=true, and the description adds behavioral context by specifying the ordering (descending by publication date). This goes beyond the annotation, providing useful execution behavior without contradiction.
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 extremely concise: two sentences that front-load the purpose and ordering. Every sentence adds value without redundancy.
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?
Given the tool's simplicity and the presence of an output schema, the description is sufficiently complete. It covers purpose, ordering, and board grouping. Missing details like pagination are not critical for a list tool with a limit parameter.
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 schema already covers 75% of parameters with descriptions. The tool description does not add additional meaning beyond implying board grouping and date ordering. It fails to clarify the 'limit' parameter or provide extra context for usage.
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 clearly states the tool lists recent documents by board in descending order of publication date, specifying the resource and action. It distinguishes from siblings like search tools by focusing on monitoring recent items.
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 implies use for monitoring recent documents but does not explicitly state when to use this tool versus alternatives like search or extract tools. No exclusions or alternative recommendations are provided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
[운영/쓰기] 게시판 크롤 → 첨부 파싱 → 적재 → 상태 재판정. 느릴 수 있음.
| Name | Required | Description | Default |
|---|---|---|---|
| board | Yes | 'm_1059'(지침서) | 'm_203'(법령) | |
| max_pages | No | 크롤할 목록 페이지 수. | |
| full | No | True면 모든 첨부 재파싱 + FTS 전체 재색인. |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description discloses behavioral traits beyond the annotations: it reveals the multi-step pipeline (crawl, parse, load, reassess) and warns about potential slowness. Annotations already indicate readOnlyHint=false, destructiveHint=false, idempotentHint=true, and the description aligns without contradiction. The additional pipeline context helps the agent understand the cost and side effects of invoking this 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?
The description is very concise: a single sentence in Korean that front-loads the operation type and pipeline steps. Every word serves a purpose, and the warning about slowness is useful. It sacrifices some structural clarity (e.g., no bullet points or emphasis) but is efficient for the information conveyed.
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?
Given the tool's complexity (multi-step refresh, multiple parameters, and output schema), the description together with annotations and schema provides sufficient context. The pipeline steps, write nature, and speed warning cover most operational aspects. However, it does not explain the purpose of the refresh (e.g., updating cached data) or when the output schema might be relevant, but the output schema itself handles return value details.
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 input schema already has full coverage (100%) with descriptions for all three parameters ('board', 'max_pages', 'full'). The description text adds no additional meaning or context about these parameters beyond what the schema provides. According to the guidelines, when schema coverage is high, the baseline is 3, and the description does not exceed this 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 explicitly defines the tool as a refresh pipeline: '게시판 크롤 → 첨부 파싱 → 적재 → 상태 재판정' (crawl bulletin board, parse attachments, load, reassess status). The verb is implied by the name 'refresh' and the pipeline. It clearly distinguishes from sibling tools like 'mfds_get_document' (single retrieval) or 'mfds_search_guidelines' (search), as it is a batch update operation.
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 provides some usage context by labeling it as '운영/쓰기' (write operation) and warning it may be slow ('느릴 수 있음'). This implies it should be used for bulk updates rather than quick queries. However, it does not explicitly state when to use this tool versus specific alternatives, nor does it mention when not to use it. The guidelines are therefore implicit rather than explicit.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
식약처 공무원 지침서/민원인 안내서(m_1059) 전문검색. 현행 추정본 우선.
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | 검색어(한국어 부분일치). 예) "의약품 심사", "복합제 독성시험" | |
| domain | No | 'drug'|'food'|'device'|'cosmetic' 필터(선택). | |
| current_only | No | True면 현행/unknown 만(폐지·구버전 제외). | |
| include_superseded | No | True면 구버전 포함. | |
| limit | No | 최대 건수. |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations declare readOnlyHint=true, and the description does not contradict it. It states the tool performs a search, which is read-only. No additional behavioral details (e.g., rate limits) are needed for a simple search 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?
Description is two short sentences, front-loads the main purpose, and contains no fluff. Could be slightly more structured, but efficient.
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?
Given the tool's complexity (5 parameters, output schema, readOnlyHint), the description is adequate. It specifies the document type and prioritization, which complements the schema. No major 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 coverage is 100%, so baseline 3. The description adds no significant meaning beyond what each parameter's description already provides in 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?
Description clearly states it is a full-text search for MFDS guidelines, specifying document type (m_1059) and prioritization of current versions. However, it does not explicitly differentiate from sibling search tool mfds_search_law_notices.
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?
Description implies use for searching guidelines but provides no guidance on when to use alternatives or exclusions. No mention of when not to use this tool.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
식약처 법령 '개정공포' 검색(m_203). ⚠ 현행 법령 본문이 아님.
각 항목은 law.go.kr 로 연결되는 개정공포 알림이며, 공포번호(anc_no)·공포일(anc_date)· 시행일(ef_date)·law_go_kr_url 을 제공한다. 법령의 현행 시행 여부(law_currency)는 항상 'unknown' 으로 반환하므로 반드시 law_go_kr_url 원문으로 확인하라.
| Name | Required | Description | Default |
|---|---|---|---|
| query | No | 검색어(법령명/키워드). 예) "식품위생법" | |
| law_name | No | 법령명 정확 필터(선택). | |
| since_date | No | 'YYYY-MM-DD' 이후 게시물만(선택). | |
| limit | No | 최대 건수. |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true. The description adds valuable behavioral context: the tool returns amendment notices (not current law), the law_currency field is always 'unknown', and the results link to law.go.kr. This goes beyond the annotation by disclosing limitations and required user verification.
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 extremely concise: two short Korean sentences plus a warning. It front-loads the purpose and key limitation, with no unnecessary words or repetition. Every sentence adds value.
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?
Given the presence of an output schema and 100% schema coverage for parameters, the description adequately covers the tool's purpose, warning, and key return fields. It mentions the critical always-unknown field. It is slightly lacking in terms of pagination details, but for a search tool this is sufficient.
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 documents all parameters (query, law_name, since_date, limit). The description does not add additional parameter-level meaning 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?
The description clearly states the tool searches for MFDS law amendment promulgation notices, explicitly distinguishing it from current law text. It specifies what each result item provides (announcement number, dates, URL). This differentiates it from siblings like mfds_latest_version and mfds_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?
The description warns that this is not current law and that the law_currency field is always 'unknown', urging users to check the original URL. While it does not explicitly list alternatives or when-not-to-use, the warning implicitly guides the agent away from using this for current law queries.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
진단: board별 문서 수·상태 분포·첨부 파싱 실패·마지막 크롤 시각·구조 경고.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The annotations already declare readOnlyHint=true. The description adds value by listing the specific types of diagnostic data returned (e.g., attachment parsing failures, structure warnings). However, it does not disclose any potential side effects, rate limits, or other behavioral traits beyond what annotations provide.
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, compact sentence that lists the output categories. It is front-loaded with '진단:' to indicate purpose. While efficient, the list format could be more structured (e.g., bullet points) for easier parsing.
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 description is marginally adequate given the presence of an output schema and no parameters. However, it is very terse and does not explain the context (e.g., what 'board' refers to) or the structure of the output. An agent might need more context to fully understand the returned data.
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?
There are zero parameters, and the schema coverage is 100%. The description does not need to add parameter semantics. The baseline for 0 params is 4, and the description does not detract.
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 clearly states the tool provides diagnostic information: document counts, status distribution, attachment parsing failures, last crawl time, and structure warnings. It specifies the resource (system status) and the action (diagnosis). However, it does not explicitly differentiate from sibling tools, and the purpose could be more action-oriented.
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 lacks any guidance on when to use this tool versus alternatives. It does not mention prerequisites, exclusions, or typical scenarios. The agent is left to infer usage from the context alone.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
특정 문서가 속한 버전 그룹의 전체 판본·상태·판정근거(audit 뷰).
| Name | Required | Description | Default |
|---|---|---|---|
| board | Yes | 'm_1059' | 'm_203' | |
| seq | Yes | 문서 seq |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description adds context beyond the readOnlyHint annotation by specifying that the tool returns '판본·상태·판정근거' (version, status, judgment basis), indicating an audit-focused read operation. However, no additional behavioral traits (e.g., rate limits, data freshness) are disclosed.
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 sentence that directly conveys the tool's purpose without wasted words. It is appropriately sized and front-loaded.
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?
Given that an output schema exists, the description need not detail return values, but it still specifies the key outputs (version, status, judgment basis). This provides sufficient context for a read-only audit tool, though it could mention any limitations or prerequisites.
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 documents both parameters (board and seq) adequately. The description does not add further parameter semantics, meeting the baseline expectation.
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 clearly identifies the tool's output (full version, status, judgment basis) and distinguishes it from sibling tools like mfds_latest_version by specifying an 'audit view' of a version group. However, it lacks an explicit action verb (e.g., 'Retrieve'), though the purpose is still clear.
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?
No guidance is provided on when to use this tool versus siblings. The description does not mention alternatives, prerequisites, or exclusions, leaving the agent to infer usage from the tool's name and output description.
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.
v0.1.0mfds_extract_sectionmfds_get_documentmfds_latest_versionmfds_list_recentmfds_refreshmfds_search_guidelinesmfds_search_law_noticesmfds_statusmfds_version_historyScored across 9 tools
Each tool targets a distinct operation: extracting sections, retrieving documents, finding latest versions, listing recent updates, refreshing data, searching guidelines, searching law notices, checking status, and viewing version history. There is minimal overlap.
All tools share the 'mfds_' prefix, but the naming pattern is inconsistent: some use verb_noun (extract_section, get_document, search_guidelines), while others use noun phrases (latest_version, status, version_history) or verb_adj (list_recent). This could confuse an agent expecting a uniform pattern.
With 9 tools, the server is well-scoped for the domain of MFDS document and law notice access. The count feels appropriate without being overwhelming or too sparse.
The server covers key reading and searching operations, but has notable gaps: there is no tool to retrieve actual current law text (only amendment notices with 'unknown' currency) and no browsing or listing of all documents beyond recent ones. Core workflows are supported but incomplete.
Korean listed medical AI, biotech and device companies linked to MFDS and FDA regulatory records.
Full-text search over K-IFRS/K-GAAP standards and KASB accounting Q&A for Korean accountants
Full-text search over FSS/FSC accounting supervision documents for Korean accounting professionals
Find official Korean public datasets, agency-site menus, disclosure listings, and source URLs.
