CITATION — REFERENCE ENTRY
Context Poisoning — Confident AI
- Key
- deepteam-context-poisoning
- Authors
- DeepTeam by Confident AI
- Type
- webpage
- Publisher
- Confident AI
Raw CSL JSON
{
"URL": "https://www.trydeepteam.com/docs/red-teaming-agentic-attacks-context-poisoning",
"type": "webpage",
"title": "Context Poisoning",
"author": [
{
"literal": "DeepTeam by Confident AI"
}
],
"accessed": {
"date-parts": [
[
2026,
5,
23
]
]
},
"language": "en",
"publisher": "Confident AI"
}
Claims
-
DeepTeam distinguishes context poisoning from direct prompt injection by noting that context poisoning does not issue instructions, but instead reframes the operational reality in which the request is answered by injecting false background context that appears as pre-existing memory, environmental state, policy precedent, or historical assumptions.
"Unlike direct prompt injection, Context Poisoning does not issue instructions. Instead, it reframes the operational reality in which the request is answered."
Available in