qf-studio/pilot

What it solves

Pilot is an AI software engineer that automates the entire lifecycle of ticket resolution. It eliminates the manual effort of picking up backlog items, planning implementations, writing code, and running tests for common tasks like bug fixes, small features, and refactoring.

How it works

Pilot integrates with project management tools (GitHub, Linear, Jira, Asana) and communication platforms (Telegram). When a ticket is assigned via a specific label or message, Pilot claims the task, creates a dedicated git branch, analyzes the codebase to plan a solution, implements the code using Claude Code, validates the changes through quality gates (tests, linting, and builds), and finally opens a pull request for human review.

Who it’s for

It is designed for developers and engineering teams who want to automate the "quick fixes" and routine maintenance tasks in their backlog to focus on more complex architectural work.

Highlights

• Multi-platform Integration: Supports GitHub, GitLab, Azure DevOps, Linear, Jira, and Asana for task ingestion.
• Autonomous Execution Modes: Offers dev, stage, and prod environments to control the level of autonomy, ranging from auto-merging to requiring human approval.
• Intelligent Routing: Automatically routes tasks between different Claude models (Haiku for trivial tasks, Opus for complex ones) based on detected effort.
• Comprehensive Monitoring: Includes a TUI dashboard for real-time tracking of queue depth, token usage, and costs.
• Interactive Telegram Bot: Allows users to chat, research codebase patterns, plan features, and execute tasks via natural language, including voice and image support.
• Cross-Project Memory: Shares learned patterns and context across different repositories to improve consistency.