External knowledge integration

The process of augmenting large-language-models with data from outside their original training set to enhance accuracy and utility within agentic-ai.

context-engineering

The art and science of optimizing the context-window by strategically injecting the necessary information at each step of an Agent’s trajectory.

• Core Components:
  • Instructions (System Prompts).
  • External knowledge (e.g., via rag).
  • tool-feedback and execution outputs.
• Implementation: Frameworks like LangGraph provide structured support for managing these context-shifting strategies.

---

Backlinks:

• 2026 04 14 Langchain context engineering