1 articles with this tag
ClawGuard offers a deterministic runtime security framework to prevent indirect prompt injection in LLM agents by enforcing user-confirmed rules at tool-call boundaries.
