
benign-instruction-bench
Re-evaluating prompt-injection detectors on LLM agent tool outputs (paper draft, scripts, scores)
ai-securitydefensive-toolseducation+3

Re-evaluating prompt-injection detectors on LLM agent tool outputs (paper draft, scripts, scores)

Research implementation of Hop-Decayed Influence (HDI) and the 3S attack framework, exposing structural auxiliary indexing vulnerabilities in…