Skip to main content
Glama

Belgeyi bu bilgisayarda maskeleyip aç

mask_file
Idempotent

Masks Turkish KVKK personal data in legal documents locally, returning only placeholder text such as [AD SOYAD-1] so sensitive details never reach the model.

Instructions

Katip KVKK kapsamındaki kişisel verileri bu bilgisayarda çıkarır; Claude, Katip'in bulduğu değerlerin yerine yer tutucuları görür. .docx, .udf (UYAP), .txt, .pdf (taranmış dahil) ya da fotoğrafı bu bilgisayarda açar, taramaları yerel OCR ile okur, kişisel verileri (ad soyad, T.C. kimlik / YKN / pasaport no, adres, telefon, IBAN, sağlık ve diğer KVKK m.6 verileri) yerelde çıkarır ve yalnızca maskeli metni döndürür. Uzun belgeler sayfa sayfa gelir.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "bilgi": "<bilgi>"
    },
    "name": "mask_file"
  }
}

Input Schema

NameRequiredDescriptionDefault
pageNo
pathNoAbsolute path, if no file_id
belgeNolist_files'taki kısaltılmış ad (ör. "… - Vatandaşlık Dava Dilekçesi.udf"); onay kutusunda kullanıcıya gösterilir
bilgiYesAlways send exactly this text; it tells the user what Katip does.Belge bu bilgisayarda açılır ve maskelenir; Claude, Katip'in bulduğu kişisel verilerin yerine [AD SOYAD-1] gibi yer tutucuları görür. Katip bazı değerleri kaçırabilir.
matterNoMatter (dosya) label; files of the same case share placeholders. Use a neutral label such as a file number, not a client name.default
file_idNoId from list_files (preferred: keeps the file name private)
profileNoai = Yapay zekâ ile paylaşım (default: all personal data, the whole address, case and bar numbers); court = Mahkemeye sunum (parties' own addresses, UETS and bar numbers masked; the disputed property's address, case numbers and dates stay readable); anon = Tam anonim (also dates, provinces, courts, companies and institutions).ai

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A4.2/5.0
Behavior5/5

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

Beyond the annotations (readOnlyHint=false, idempotentHint=true, destructiveHint=false, openWorldHint=false), the description discloses the critical privacy model: extraction happens on this machine, Claude only sees placeholders like [AD SOYAD-1], and only masked text is returned. It also warns that Katip may miss some values and notes long documents arrive page by page, which is exactly the behavioral context an agent needs.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The key point — local extraction with placeholders — is front-loaded in the first sentence, and every later clause adds concrete facts (formats, OCR, data categories, pagination). It loses a point for repetition of 'bu bilgisayarda'/'yerelde' three times and a dense run-on layout without breaks.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 7-parameter tool with no output schema, the description covers what the tool does, what the caller receives (masked text, page by page), supported file types, OCR handling of scans, PII categories, and the reliability caveat. Nothing essential for correct invocation is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 86% schema coverage the schema already carries most parameter meaning. The description contributes indirectly by explaining that long documents come page by page (giving meaning to the 'page' parameter) and that partial results are possible, but it does not clarify 'bilgi', 'profile', 'matter', or 'file_id' beyond what the schema strings state. Baseline 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description names a concrete verb chain (open locally, OCR, extract KVKK personal data, return only masked text) plus the exact supported formats (.docx, .udf, .txt, .pdf, photo). The clause 'yalnızca maskeli metni döndürür' implicitly separates it from siblings that write results back to disk (write_masked_copy, restore_to_file). An agent can identify the operation without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Usage is implied by the data types listed (personal data under KVKK) and the fact that only masked text comes back, so an agent can infer this is the tool for reading documents containing PII. However, no sibling is named and there is no explicit when-not-to-use or prerequisite guidance (e.g. when to prefer write_masked_copy or restore_to_file).

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