Loading content
Context poisoning describes contamination of retrieval sources, memory, instructions, or work artifacts. An agent adopts false assumptions and may propagate them across multiple work steps.
Use this profile to understand the building block briefly, place it in the model, and open related building blocks.
Theoretical construct: explains a term, principle, or mental model.
Concrete cog in the system that works inside larger relationships.
Context poisoning manipulates examples, documents or memories supplied to an AI agent so that it later makes harmful decisions.
The approach transfers data and model poisoning, including contextual backdoors, to external agent contexts. Research on embodied agents showed in 2024 how a few poisoned demonstrations can change behaviour under selected triggers.
The agent reads a tampered manual: it appears plausible but steers the agent to a wrong action in the right situation.
Attackers alter the working context rather than the model and influence downstream actions.
Provenance, validation, permissions and logging limit the impact.
Context poisoning shows why external content and examples must be treated as security-sensitive inputs.
Where this building block is located in the topic model.
No structure path available.
Explore how this building block connects to concepts, methods, technologies, and tools.
These sources establish the term and its professional meaning.
All direct connections of the current building block in a compact text view.
This classification shows where the building block typically matters, how demanding it is, and what kind of impact it has in the model.
The level within the organization (enterprise, domain, team) at which the AssetBlock is applied.