auto-bug-fix: An AI-Native Automated Bug Fix Scheduling System

auto-bug-fix is an AI-native automated bug fix scheduling system that connects Jira and GitLab. It automatically assigns bugs to AI agents to perform analysis, fixing, testing, and submitting merge requests. Its core scheduler is implemented in Go and supports integration with multiple AI agents (such as OpenAI Codex, Claude Code, etc.), aiming to solve the problem of integrating AI tools with existing enterprise development workflows and improve bug fix efficiency.

In software development, developers spend an average of 30%-40% of their time on bug fixing, which affects efficiency and easily leads to burnout. After the rise of AI coding assistants, enterprises hope to integrate them into their development processes, but there is a lack of reliable scheduling and coordination for automated AI-driven bug fix processes between systems like Jira (issue tracking) and GitLab (code hosting). The auto-bug-fix project was born to address this, focusing on the automatic assignment and execution of bug fix tasks in enterprise environments.

auto-bug-fix uses a two-layer architecture:

1. Go Core Scheduling Layer: Responsible for polling Jira to get bugs, priority sorting and assignment, managing workflow state machines, interacting with GitLab (creating branches/merge requests), logging, and other deterministic tasks, emphasizing reliability and predictability.
2. AI Agent Layer: Supports multiple AI tools such as OpenAI Codex, Kiro, Claude Code, Cursor, etc., to complete fixing tasks. Its agent-agnostic design allows flexible switching of AI tools, parallel processing of the same bug by multiple agents, and reduces dependency on a single vendor.

The workflow of auto-bug-fix is divided into three stages: 1. Bug Discovery and Evaluation: Poll Jira to get bug tickets → intelligently filter bugs suitable for automatic fixing → evaluate complexity and code change scope to decide whether to assign to AI. 2. AI Fix Execution: Create an isolated code branch → collect context (code, test cases, etc.) → schedule AI agents to analyze the bug and generate fix code → run tests for verification. 3. Quality Verification and Submission: Automatically run unit/integration tests → optional manual review → create merge request in GitLab → sync results to Jira.

1. Go Language Core: Uses goroutines for concurrent task processing, easy deployment as a single binary, static type safety, and a rich ecosystem supporting Jira/GitLab integration.
2. Extensible Skill System: Defines processing strategies (context collection, AI agent selection, test verification, etc.) for different bug types (security vulnerabilities, performance issues, etc.).
3. Security and Isolation: AI agents run in sandbox containers, minimal permissions, detailed audit logs, and manual confirmation at key nodes.

auto-bug-fix is suitable for:

• Large Codebase Maintenance: Automatically handle low-priority bugs, allowing teams to focus on new features.
• Open Source Project Maintenance: Reduce the burden on maintainers by automatically handling simple bugs.
• Security Vulnerability Response: Quickly generate fix patches to shorten the vulnerability exposure window.
• Code Quality Improvement: Automatically fix issues like code style and performance anti-patterns.

Limitations: AI struggles to handle bugs involving complex business logic or architectural refactoring; relies on project test coverage; has limited understanding of the overall project architecture; has security and compliance risks. Future Directions: Enhance learning capabilities (improve from human feedback); multi-agent collaboration to handle complex bugs; intelligent routing of bugs to appropriate AI agents; integrate platforms like GitHub Issues/Azure DevOps; predict fix effectiveness.

auto-bug-fix represents an important direction in AI-assisted development—deeply integrating AI into enterprise workflows. By separating deterministic scheduling and intelligent fixing, it balances reliability and flexibility. Although it cannot completely replace human developers, it can significantly improve efficiency when handling simple and repetitive bugs. Its open-source nature allows community participation in improvements, and it will play a more important role in software development in the future, providing teams with a solution to improve efficiency and reduce maintenance burdens.