<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
        xmlns:news="http://www.google.com/schemas/sitemap-news/0.9">
  <url>
    <loc>https://samtaelabs.com/blog/claude-code-vs-codex-it-agency-guide</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-02T22:30:02.990Z</news:publication_date>
      <news:title>클로드 코드 vs 코덱스, IT 에이전시는 어떤 AI 코딩 도구를 선택해야 하나</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/ai-agent-architecture-for-dev-agency</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-02T12:30:35.920Z</news:publication_date>
      <news:title>AI 에이전트, 어떻게 설계해야 실제로 작동하는가 — IT 개발 에이전시가 봐야 할 구조</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/uber-claude-code-ai-budget-exhausted-2026</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-02T12:30:35.860Z</news:publication_date>
      <news:title>Uber가 4개월 만에 2026년 AI 예산을 소진한 이유 — Claude Code 도입이 남긴 교훈</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/ai-agent-harness-engineering-for-dev-agency</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-02T03:00:27.854Z</news:publication_date>
      <news:title>AI 에이전트를 외주 프로젝트에 붙이기 전에 반드시 알아야 할 것: 하네스 엔지니어링</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/ai-agent-era-it-agency-three-things-to-fix</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-01T22:30:02.994Z</news:publication_date>
      <news:title>AI 에이전트 시대, IT 에이전시가 클라이언트 프로젝트에서 놓치고 있는 세 가지</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/ai-agent-harness-design-for-development-agency</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-01T22:30:02.776Z</news:publication_date>
      <news:title>AI 에이전트 개발, 모델보다 하네스 설계가 성패를 가른다</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/why-ai-coding-fails-at-product-level</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-01T12:30:35.957Z</news:publication_date>
      <news:title>AI가 코드는 짜도 제품은 못 만드는 이유 — 개발 에이전시 관점</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/ai-library-supply-chain-attack-risk-for-dev-agency</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-05-01T12:30:35.892Z</news:publication_date>
      <news:title>AI 개발 라이브러리에 악성코드가 숨어 있었다면? 공급망 공격과 외주 개발의 리스크</news:title>
    </news:news>
  </url>
</urlset>