Skip to content
This repository was archived by the owner on May 17, 2025. It is now read-only.

gsoc2025_ideas.md

Latest commit

 

History

History
111 lines (71 loc) · 3.87 KB

gsoc2025_ideas.md

File metadata and controls

111 lines (71 loc) · 3.87 KB
layout default
title Rspamd project ideas for GSOC 2025

Rspamd project ideas for GSOC 2025

Introduction

This page lists project ideas for contributors to Rspamd, particularly for programs like Google Summer of Code (GSoC). We welcome contributors who are passionate about open-source development, email/spam filtering, and modern technologies like AI and Rust.

Information for GSoC Participants

Prospective contributors should:

  • Have a GitHub account and familiarity with Git workflows.
  • Review the Rspamd repository.
  • Join our Telegram channel rspamd for discussions.
  • Demonstrate proficiency in C (or C++), Lua or Rust (depending on the project).

All code must be licensed under Apache 2.0.

Mentors

Mentor Email Role
Vsevolod Stakhov vsevolod@rspamd.com Core Development, Admin
Andrew Lewis alewis@rspamd.com Core Development, Lua plugins development
Anton Yuzhaninov citrin@rspamd.com Protocols, Integrations

List of Projects

Multi-Class Bayesian Classifier

Description: Extend Rspamd’s Bayesian classifier to support multiple categories (beyond spam/ham) and integrate AI-driven learning via a GPT plugin for dynamic model updates.

Difficulty: Medium/Hard

Timeline: 22 weeks

Skills: Machine Learning (Bayesian methods), Lua, Python (for GPT integration)

Mentors: Vsevolod Stakhov, Andrew Lewis

Benefits: Gain expertise in AI/ML integration, probabilistic classifiers, and large-language model APIs.
Evaluation:

  • Midterm: Basic multi-class support in Bayes module; GPT plugin prototype.
  • Final: Full integration with GPT for automated learning; performance benchmarks.

Full Telegram Support (Bot for Spam Filtering)

Description: Implement integration of with Telegram bot for spam filtering, including rule-based automation (e.g., user reports, admin moderation).

Difficulty: Medium

Timeline: 12 weeks

Skills: Rust (Telegram Bot API), Lua, Rule Engine Design

Mentors: Andrew Lewis, Anton Yuzhaninov

Benefits: Learn real-time bot development, protocol integration, and spam rule optimization.
Evaluation:

  • Midterm: Functional Telegram bot with basic spam reporting.
  • Final: Advanced rules (e.g., rate limiting, user reputation), moderation UI, and documentation.

Settings Manager (UI + Rust Backend)

Description: Build a user-friendly UI for managing Rspamd settings and a Rust-based backend for storing configurations in MySQL/PostgreSQL.

Difficulty: Medium

Timeline: 12 weeks

Skills: Rust, JavaScript/TypeScript (React/Vue), SQL

Mentors: Andrew Lewis, Vsevolod Stakhov

Benefits: Master full-stack development, Rust database integration, and secure UI design.
Evaluation:

  • Midterm: Rust backend with CRUD operations; UI prototype.
  • Final: Full UI feature set (import/export, versioning), performance optimizations.

GnuPG Signing and Verification Support

Description: Enhance Rspamd’s GnuPG support for signing/verifying emails, including key management and policy enforcement.

Difficulty: Hard

Skills: C, Cryptography (PGP/GnuPG), Lua

Timeline: 22 weeks

Mentors: Vsevolod Stakhov

Benefits: Deepen knowledge of cryptographic protocols and secure C programming.
Evaluation:

  • Midterm: Basic message signing/verification workflow.
  • Final: Key rotation policies, WebUI integration, and attack-resistance testing.

How to Proceed

  1. Fork the Rspamd repo and explore the codebase.
  2. Discuss your proposal with mentors on Telegram or the mailing list.
  3. Submit a detailed timeline with milestones matching GSoC’s 12-week schedule.

We value passion, clarity, and realistic planning!