UTC

CWE-1427 · Base

Improper Neutralization of Input Used for LLM Prompting

The definition on cwe.mitre.org →
OWASP LLM Top 10 2026 risks behind it
  • ●LLM01:2026 Prompt InjectionUntrusted input from prompts, retrieved documents, tool output, or memory alters model behavior because the model draws no boundary between instructions and data, the definitional prompt-injection weakness.
  • ○LLM03:2026 Excessive AgencyDirect and indirect prompt injection, such as a crafted email, is the trigger that turns excess permissions into harmful action.
  • ○LLM05:2026 Data and Model PoisoningPoisoned retrieval documents and hidden web instructions reach the prompt without neutralization, the indirect-injection facet of poisoning.
  • ○LLM07:2026 MisinformationAdversarially crafted inputs induce false or misleading output, the input-side manipulation behind adversarially induced misinformation.

From OWASP's Appendix A, under CC BY-SA 4.0. ● primary · ○ supporting.