Skip to content

mbtariq82/bug-agent

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Bug Agent

An autonomous agent to help maintainers research issues/bugs. Potential workflows include the following:

  • searching through existing issues/PRs
  • recreating a bug
  • writing a comment (see github.com/huggingface/transformers/issues/44485 → references to vllm and sglang)
  • summary of a PR for a specific maintainer

Pipeline:

  1. GitHub API
  2. Issue Summarizer / PR Summarizer
  3. LM advisor
  4. SurrealDB for persistent memory

Quick Start

(Optional) Change the model used by the advisor. Default is HuggingFaceTB/SmolLM-1.7B:

export MODEL_NAME=HuggingFaceTB/SmolLM-1.7B
# Windows PowerShell
$env:MODEL_NAME = 'HuggingFaceTB/SmolLM-1.7B'

Run the agent (defaults to the latest open Transformers issue):

python -m src.main
  • To process a specific issue:
python -m src.main --issue 44593

About

A tool to research GitHub issues. Using SurrealDB for persistent memory, the agent is capable of building a knowledge base for the repo. Bug-Agent helps open source maintainers search through existing issues/PRs, recreate bugs, write and review PRs.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages