StavC/Here-Comes-the-AI-Worm
Here Comes the AI Worm: Preventing the Propagation of Adversarial Self-Replicating Prompts Within GenAI Ecosystems
Implements RAGworm, a self-replicating prompt injection attack that propagates across RAG-connected GenAI applications with super-linear growth (20 new compromises per client in 1-3 days). DonkeyRail, the accompanying guardrail, detects and blocks these worms using machine learning with 1.0 true-positive rate and minimal 7.6-38.3ms latency overhead. Includes datasets, evaluation frameworks, and end-to-end demos targeting email assistants and productivity tools.
222 stars. No commits in the last 6 months.
Stars
222
Forks
27
Language
Jupyter Notebook
License
—
Category
Last pushed
Sep 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/prompt-engineering/StavC/Here-Comes-the-AI-Worm"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
liu00222/Open-Prompt-Injection
This repository provides a benchmark for prompt injection attacks and defenses in LLMs
cybozu/prompt-hardener
Prompt Hardener analyzes prompt-injection-originated risk in LLM-based agents and applications.
R3dShad0w7/PromptMe
PromptMe is an educational project that showcases security vulnerabilities in large language...
lakeraai/pint-benchmark
A benchmark for prompt injection detection systems.
grepstrength/WideOpenAI
Short list of indirect prompt injection attacks for OpenAI-based models.