# 4 CREWAI - JOB HUNTER AGENT # 목차 - [4.0 Introduction](#40-introduction) - [4.1 Agents and Tasks](#41-agents-and-tasks) - [4.2 Context And Structured Outputs](#42-context-and-structured-outputs) - [4.3 Firecrawl Tool](#43-firecrawl-tool) - [4.4 Knowledge Sources](#44-knowledge-sources) - [4.5 Conclusions](#45-conclusions) # 4.0 Introduction # 4.1 Agents and Tasks # 4.2 Context And Structured Outputs ## Context ```python @CrewBase class JobHunterCrew: ... @task def interview_prep_task(self): return Task( config=self.tasks_config["interview_prep_task"], # context: List of Task instances # providing task context or input data. context=[ self.job_selection_task(), self.resume_rewriting_task(), self.company_research_task(), ] ) ``` ## Model(Schema) - 출력 구조를 강제하면 더 나은 결과를 얻을 수 있음 ```python from pydantic import BaseModel class Job(BaseModel): job_title: str company_name: str job_location: str is_remote_friendly: bool | None = None employment_type: str | None = None compensation: str | None = None job_posting_url: str job_summary: str ... ``` # 4.3 Firecrawl Tool # 4.4 Knowledge Sources # 4.5 Conclusions