Skip to main content
Glama
Sign In

Æfact — Korean fact desk for AI agents

by site.aefact
OverviewDiscussionsAdmin

AI 팩트체크

fact_check
Read-onlyIdempotent

글(답변 초안·기사·보고서, 4000자까지)을 통째로 넘기면 팩트체크가 필요한 문장을 골라 대신 확인합니다. 법령 시행일·한국은행 기준금리·월별 소비자물가 상승률은 원자료와 대조하고, 문장 안 증감률·비율·합계는 다시 계산합니다. 문장마다 status(확인함·사람 확인 필요·원자료 미연결·확인 대상 아님·다루지 않음)와 계산·근거 주소가 오고, attention 에 손봐야 할 문장 번호가 옵니다. 판정은 규칙·계산이며 모델을 쓰지 않습니다. 글은 저장·공개하지 않습니다. EN: Hand over a whole Korean text; Æfact picks the sentences that need fact-checking and verifies them against primary data (law effective dates, BOK base rate, monthly CPI) or by recomputing in-sentence arithmetic. Returns per-sentence status, calculations and evidence URLs. Rule-based, no LLM. Text is not stored.

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": {
      "text": "<text>"
    },
    "name": "fact_check"
  }
}

Input Schema

NameRequiredDescriptionDefault
textYes확인할 글. 문장이 여러 개여도 됩니다.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Added

TDQS

A4.4/5.0
Behavior5/5

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

Beyond the annotations (readOnly, idempotent, openWorld, non-destructive), the description discloses that judgments are rule/calculation-based with no model involved, that the text is neither stored nor published, and that results include per-sentence status values plus an attention list. This is meaningful behavioral context an agent could not infer from the 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.

Conciseness4/5

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

The Korean text is front-loaded and information-dense: purpose, coverage, return shape, determinism, and privacy each appear once in a single efficient paragraph. The full English restatement duplicates the same content and roughly doubles the length, which is the main structural cost.

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?

There is no output schema, yet the description fully specifies the return contract: per-sentence status (with the five enumerated values), calculations, evidence URLs, and an attention list of sentence numbers. Combined with the stated input limits and the no-storage guarantee, nothing an agent needs to call it correctly is missing.

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

Parameters4/5

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

With a single fully documented parameter, the schema already carries most semantics; the description still adds that the entire multi-sentence text should be passed wholesale and restates the 4000-character limit. It does not add format guidance beyond that, so it exceeds the baseline only slightly.

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 states a specific verb and resource: it takes a whole Korean text (up to 4000 chars), selects the sentences needing verification, and checks them. It names the concrete data sources (statute effective dates, BOK base rate, monthly CPI) and the arithmetic re-computation, which distinguishes it from sibling retrieval tools like get_fact, lookup, and request_check.

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 rather than stated: the agent learns to pass an entire draft/article/report rather than single claims, but the description never explicitly says when to prefer this over lookup, get_fact, or request_check, nor when not to use it. The 4000-char ceiling is the only concrete boundary given.

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

All Æfact — Korean fact desk for AI agents toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources