agent_learning

This repository is a collection of learning and portfolio projects.

Projects

• agent_rag: local academic Agentic RAG system built with Python, LangGraph, FastAPI, hybrid retrieval, Web/Scholar tools, evaluation, and harness skills.
• ai_coding: local AI coding agent CLI built with Python and LangGraph, covering workspace-scoped edits, command permissions, session recovery, structured trace, evidence cache, and harness skills.
• atomic_agents: atomic agent orchestration runtime with approval gates, retries, scheduling, and multi-agent orchestration templates.

Skills

• goal-orchestrator: goal contracts, milestones, steering, handoffs, and completion audits for long-running agent work.

Each project lives in its own directory so dependencies, runtime files, tests, and documentation can evolve independently.

See agent_rag/README.md or agent_rag/README.zh-CN.md for setup, usage, and quality checks.

See ai_coding/README.md or ai_coding/README.zh-CN.md for the AI coding agent CLI.