H3XAOperational intelligence
Live analytical pictureEvidence before assessment
Threat intelligence / Technique

Special Character Sets

Adversaries may discover delimiters and special characters sets used by the large language model. For example, delimiters used in retrieval augmented generation applications to differentiate between context and user prompts. These can later be exploited to confuse or manipulate the large language model into misbehaving.

Technique
TypeTechnique
External ID—
First seen2026-08-14
Last seen2026-08-14

Confidence

—

Adversaries may discover delimiters and special characters sets used by the large language model. For example, delimiters used in retrieval augmented generation applications to differentiate between context and user prompts. These can later be exploited to confuse or manipulate the large language model into misbehaving.

6 recorded

Relationships

mitigates
from Control Access to AI Models and Data at Rest
Mitigation
mitigates
from Control Access to AI Models and Data in Production
Mitigation
mitigates
from Generative AI Guardrails
Mitigation
subtechnique-of
to Discover LLM System Information
Technique
uses
from Financial Transaction Hijacking with M365 Copilot as an Insider
Campaign
uses
from OpenClaw Command & Control via Prompt Injection
Campaign
1 external

References

mitre-atlas
AML.T0069.000
open
Continue into the dataIntelligence explorerHow to read the graphEvidence & provenance