# generation
7 articlestagged with “generation”
CI/CD Code Generation Risks
Security risks when AI generates code executed in CI/CD pipelines with elevated privileges.
CI/CD Code Generation Risks (Code Gen Security)
Security risks of AI-generated code in CI/CD pipelines including automated merge attacks, test generation manipulation, and pipeline injection.
Advanced Test Generation Manipulation
Advanced techniques for manipulating AI-generated tests to create false assurance by generating tests that pass but don't verify security properties.
Synthetic Data Poisoning
Attacking synthetic data generation pipelines to produce poisoned training sets, including generator manipulation, prompt poisoning, and contamination amplification.
Adversarial Dataset Generator
Creating tools that generate diverse adversarial datasets for benchmarking LLM safety, including semantic variations and encoding permutations.
Payload Generation Pipeline
Build an automated pipeline for generating, mutating, and testing prompt injection payloads.
Full Engagement: Content Generation Platform
Full engagement walkthrough for a content generation AI platform with brand safety and copyright concerns.