Doktor MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| DOKTOR_MCP_CACHE_DIR | No | Cache directory for the server. If provided, cache files are written there; otherwise they go to $XDG_CACHE_HOME/doktor-mcp or ~/.cache/doktor-mcp. | |
| DOKTOR_MCP_SOURCE_MODE | No | Alternative to DOKTOR_MCP_DEFAULT_SOURCE_MODE. Sets the default sourceMode. If both are set, DOKTOR_MCP_DEFAULT_SOURCE_MODE takes precedence. | |
| DOKTOR_MCP_DEFAULT_SOURCE_MODE | No | Default source mode for tool calls. Valid values: 'mock', 'live', 'snapshot'. Defaults to 'live' (since 0.59.0). Setting to 'mock' adds a mockDataWarning to responses. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| classify_medical_legal_questionA | Advanced/granular tool: prefer prepare_doctor_legal_information_pack for end-to-end questions; use this only when you specifically need medical-legal question classification in isolation. Classifies a physician question into legal dimensions (patient_rights, professional_ethics, criminal, privacy_kvkk, civil_compensation, etc.) and extracts search terms. Example: { "question": "Hasta onami olmadan mudahale edilirse ne olur?" } |
| search_health_legislationA | Advanced/granular tool: prefer prepare_doctor_legal_information_pack for end-to-end questions; use this only when you specifically need legislation search in isolation. Searches health-related legislation by keywords. Returns legislation titles, document IDs, and matched provisions. Example: { "searchTerms": ["hasta haklari", "tedaviyi red"], "sourceMode": "mock" } |
| get_legislation_provisionsA | Advanced/granular tool: prefer prepare_doctor_legal_information_pack for end-to-end questions; use this only when you specifically need legislation provisions by document ID in isolation. Fetches full text of legislation provisions by document IDs. Example: { "documentIds": ["mevzuat:hastahaklari-yonetmelik:madde-5"], "sourceMode": "mock" } |
| search_health_precedentsA | Advanced/granular tool: prefer prepare_doctor_legal_information_pack for end-to-end questions; use this only when you specifically need court precedent search in isolation. Searches health-related court decisions from Yargitay and Danistay. Returns decision summaries, document IDs, and relevance scores. Example: { "searchTerms": ["hekim sorumlulugu", "tedaviyi red"], "sourceMode": "mock" } |
| filter_reasoned_precedentsA | Advanced/granular tool: prefer prepare_doctor_legal_information_pack for end-to-end questions; use this only when you specifically need to filter/classify court decisions for reasoned analysis. Takes a list of decisions and filters them by eligibility (full text, legal reasoning, relevance). Optionally accepts a packId to operate on a previously prepared pack's decisions. Example: { "decisions": [...], "query": "hekim ihmal" } |
| prepare_doctor_legal_information_packA | PRIMARY ENTRY POINT — use this tool first for any physician legal question. Prepares a complete source-grounded legal information pack: classification, verbatim legislation provisions, verified high-court precedents, missing information, and lawyer review points. Returns a packId for follow-up drill-down. Set sourceMode:'live' to query official sources (mevzuat.gov.tr, Yargitay, Danistay); this is the default mode; sourceMode='mock' returns clearly-marked fixture data for testing only. Example: { "question": "Hasta tedaviyi reddederse hekimin sorumluluğu nedir?", "sourceMode": "live" } |
| drill_down_pack_itemA | Follow-up tool — call ONLY after prepare_doctor_legal_information_pack. Pass the packId from that response plus a follow-up question mentioning a specific article number, law name, chamber, or decision number to retrieve just that item without rebuilding the pack. Example: { "packId": "pack-3f9a2c", "followUpQuestion": "madde 24 ne diyor?" } |
| get_decision_full_textA | Fetch the full verbatim text of a single court decision by its documentId (from search_health_precedents or a pack's verifiedHighCourtPrecedents). Use for deep-dive into one decision instead of re-running the whole pack. Example: { "documentId": "yargitay:99001", "sourceMode": "live" } |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| hekim-hukuki-soru | Hekimler için yapılandırılmış hukuki soru formatı şablonu. Bu şablonu kullanarak sorunuzu daha iyi yapılandırabilirsiniz. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| health-legislation-inventory | Anayasal sağlık mevzuatı envanteri — health legislation inventory with access status, last verified dates, and official gazette issue details. |
| source-calibration-status | Kaynak kalibrasyon statüsü — canlı kaynakların erişilebilirlik durumu (önbellekli, 10 dk TTL). |
TDQS
Scored across 8 tools
Most tools have clearly distinct resource/action targets: pack preparation, pack drill-down, decision full-text retrieval, legislation search/fetch, and precedent search/filter/classification. The only mild ambiguity is that several granular tools are explicitly subsumed by prepare_doctor_legal_information_pack, so an agent may occasionally wonder whether to use the umbrella tool or an isolated component.
All tool names use consistent snake_case and follow a predictable verb_noun or verb_object pattern. The naming is verbose but uniform across the set.
Eight tools is well-scoped for a domain-specific legal information server. Each tool corresponds to a meaningful operation or drill-down path, with no apparent padding.
The surface covers the full workflow: question classification, legislation search and provision retrieval, precedent search and filtering, full decision retrieval, pack assembly, and follow-up drill-down. No obvious lifecycle gaps are present for the stated physician legal information purpose.