GSoC 2026 at OWASP BLT

NetGuardian is a zero-trust ingestion and triage pipeline that connects distributed security producers to BLT. A Cloudflare Python Worker verifies signed ztr-finding-1 envelopes using JCS and HMAC, enforces replay protection and RBAC, stores metadata in D1, keeps encrypted evidence in R2, and enriches findings through BLT-API for CVE-aware scoring and issue creation. An events outbox delivers to downstream consumers; operators review via a static triage UI. Deliverables include versioned ingestion APIs with OpenAPI, end-to-end swarm dataflow, observability and rollout gates with SLOs, and a midterm slice for performance and correctness validation.

Contributor

Preetham Poojari

Pritz395 Work product

An interactive security learning platform on OWASP BLT that turns labs into hands-on, code-driven exercises. Django-based labs migrate to a static Jekyll workflow with Markdown content. Learners follow Identify → Explain → Fix on real vulnerable code. GitHub Actions validate and sync content; a security intelligence pipeline pulls patterns from issues, PRs, and reviews into dashboards and recommendations linking real findings to labs—building practical security mindset and healthier contributions.

Contributor

Sakshee Suman

e-esakman

Full migration from the Django monolith to a static frontend on GitHub Pages with Cloudflare Python Workers at the edge. Modular independent services replace bundled Django features: static experiences ship via GitHub Pages; server logic runs on Workers with Cloudflare D1—the goal is near-zero hosting cost using free-tier Workers and Pages while preserving capability.

Contributor

Md Kaif Ansari

mdkaifansari04

Evolves BLT-Vanish into a proactive identity defense system: AI-assisted phishing detection, credential hygiene, and correlated suspicious logins. The Universal Identity Shield uses a hybrid local-first and edge architecture for low-latency, explainable risk scoring—phases cover a foundational risk engine, core workflows (dark web monitoring, guided password rotation), and production hardening with Flutter and Python Cloudflare Workers. Deliverables include phishing detection, a zero-plaintext encrypted vault, anomaly correlation, and full documentation.

Contributor

Ananya

ananya-09

Building a local-first, 3-tier AI-powered security advisory system (VS Code Extension + Pre-commit Hook + GitHub Actions) that gives contributors real-time guidance on security risks using hybrid risk classification and RAG over OWASP documentation.

Interns

Savio D'souza

S3DFX-CYBER

BLT-Toasty is a Cloudflare-native AI triage assistant that automates the processing of OWASP vulnerability reports. To maintain strict responsible disclosure standards, Toasty enforces a fail-closed PII redaction gate using Microsoft Presidio before any AI processing occurs. The system provides token-optimized LLM summarization, embedding-based duplicate detection via Cloudflare Vectorize, and structured MCP export for agent-driven automation. By delivering core infrastructure like webhook ingress, a Label Rationale API, and Durable Objects state management, Toasty targets a ≥30% reduction in manual triage time and ≥0.80 duplicate detection precision—transforming raw reports into clean, actionable security intelligence.

Interns

Ojaswa

ojaswa072

Internship project track for BLT-MCP. Detailed project scope will be published as milestones are finalized.

Interns

Nachiket Roy

Nachiket-Roy

Internship project track for SafeCloak. Detailed project scope will be published as milestones are finalized.

Interns

Karuna Rapolu

karunarapolu

Internship project track for Hackerhouse. Detailed project scope will be published as milestones are finalized.

Interns

Rudra Pratap Singh

Rudra-rps

Internship project track for BLT-Autopatch. Detailed project scope will be published as milestones are finalized.

Interns

Sidd190

Sidd190

Internship project track for BLT-MAYO. Detailed project scope will be published as milestones are finalized.

Interns

Jayant

Jayant2908