GH Repository · zhnt
loushang
AI-native agent harness for coding workflows by python: multi-model LLM orchestration, stateful sessions, tool governance, traceable delivery, and provider routing for GPT, Claude, DeepSeek, Qwen, Kimi, GLM, and MiniMax.
- stars
- 1,471
- 30-day movement
- +4515/day
- Related entries
- 60
- Connections
- 2
python/uvmakePythondynamic-workflowsagent-appagent-harnessvibe-codingaiparallel-agentsagenticagentclaude-codechatgptcodexcodingpythondeepseekglmclaudeminimaxkimiharnessworkflowskills
loushang is a Python agent harness for coding workflows, described as AI-native with multi-model LLM orchestration, stateful sessions, tool governance, and traceable delivery. It routes across providers including GPT, Claude, DeepSeek, Qwen, Kimi, GLM, and MiniMax, and ships an AGENTS.md file.
You want a single Python harness that coordinates coding agents across multiple LLM providers with session state and governed tool use.
Use it to
- Orchestrate coding agents across multiple LLM providers
- Run stateful multi-step coding sessions
- Govern and trace agent tool usage
- Build dynamic or parallel agent workflows
- Swap between GPT, Claude, DeepSeek, Qwen, Kimi, GLM, or MiniMax backends
For Python developers building multi-model coding agent workflows
- Role
- agent-app
- Language
- Python
- Licence
- Apache-2.0
- Forks
- 235
- Open issues
- 40
- Last push
- 2026-09-14
topicsagent-harnesscodingmulti-modelpythonworkflowllm-orchestration