Loading...
Loading...
Available on 1 platform
Sign in to view source links and access this dataset
1,000 to 10,000 labeled text samples for detecting indirect prompt injection attacks within repository files, released by prodnull in March 2026. The data covers READMEs, CI/CD workflows, and documentation used as context for AI coding agents.
Users should refer to the five associated Arxiv papers (e.g., 2602.14161) for methodology as column names are not explicitly documented.