YerbaPage/SWE-Debate

SWE-Debate: Competitive Multi-Agent Debate for Software Issue Resolution

29
/ 100
Experimental

Integrates the Moatless framework for graph-driven code entity extraction and dependency traversal, then employs Monte Carlo Tree Search with multiple specialized LLM agents (ReAct-based reasoning, value estimation, discriminator voting) to collaboratively debate and refine fault localization. The architecture chains entity identification through code dependency graphs, uses MCTS for exploration with multi-agent consensus voting, and generates structured code modification plans guided by feedback loops.

No Package No Dependents
Maintenance 6 / 25
Adoption 7 / 25
Maturity 9 / 25
Community 7 / 25

How are scores calculated?

Stars

25

Forks

2

Language

Python

License

Apache-2.0

Category

ai-debate-arenas

Last pushed

Nov 11, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/YerbaPage/SWE-Debate"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.