A task-agnostic LangGraph framework for Planner–Specialist–Evaluator agents: model any generate→verify→fix workflow as a state graph with a dual-gate anti-hallucination loop. Just add a task under tasks/. First task: a personal-CRM data-quality watchdog.
python workflow multi-agent watchdog data-quality verifier self-correction agent-framework llm prompt-engineering langchain state-graph langgraph anti-hallucination planner-specialist
-
Updated
Aug 9, 2026 - Python