BBomToBeinbomtobe-orke.hashnode.dev·8h ago · 2 min readA practical review loop for coding agentsAn agent can finish the code and still leave you with the hard part: proving the change matches the request. For a solo builder, review is a real task, not an afterthought. Before a run, write down on00
BBomToBeinbomtobe-orke.hashnode.dev·13h ago · 1 min read에이전트에게 일을 맡긴 다음, 결과는 어디서 확인할까요?AI 코딩 에이전트와 함께 일할 때 프롬프트만큼 결과를 확인하는 과정도 중요합니다. 코드를 수정한 뒤 테스트가 통과해도, 실제 화면이나 저장소의 변경이 의도와 맞는지 살펴봐야 합니다. 저희 BOMTOBE가 만든 orke는 Claude Code와 Codex 같은 에이전트를 위한 데스크톱 작업공간입니다. 터미널에서 일을 맡기고 같은 작업공간의 웹 창과 Git에00
BBomToBeinbomtobe-orke.hashnode.dev·17h ago · 2 min readA Run Ledger Makes Long-Running AI Work Easier to ResumeA long-running agent task can fail without crashing. A scheduled run may never start, a browser action may wait for approval, or a public action may finish while the session loses its place. A transcr00
BBomToBeinbomtobe-orke.hashnode.dev·17h ago · 3 min readA goal brief for an AI agent should define the finish lineAn instruction like “improve onboarding” gives an agent a direction, but leaves important choices open. Which users? Which part of onboarding? What evidence should be checked? What should happen if th00
BBomToBeinbomtobe-orke.hashnode.dev·1d ago · 3 min readA Completion-Time Chart Cannot Measure Agent AutonomyAn agent can finish at 2 a.m. without doing autonomous work at 2 a.m. It may have started the night before, worked for hours, waited for a person, or been launched by a schedule. A completion-time cha00
