Analysis · curated 3 Aug 2026
Prompt Injection | Giskard Documentation
First reported giskard.ai
Coverage timeline
Single-source analysis — first reported, latest, and curated coincide.
Why it matters
Prompt injection is a foundational attack class against LLM-backed systems, and reference material describing its variants and detection signals helps defenders understand where to focus hardening and testing.
A glossary/reference entry from Giskard's documentation defines prompt injection and its variants (direct, indirect, and character injection), outlines business impacts such as data leakage and content-filter bypass, and describes how Giskard's Hub UI and Python SDK can generate adversarial test cases and evaluation checks to test for the vulnerability.