Glossary

Plain-language definitions of SEO terms used across the blog.

Guardrail

A control layer around an LLM that inspects and constrains inputs and outputs to enforce safety, policy, and security boundaries.

Indirect Prompt Injection

A prompt injection delivered through external content the LLM consumes, rather than typed directly by the user.

Jailbreak vs. Prompt Injection

Two related but distinct LLM attacks: jailbreaks bypass model safety alignment, while prompt injection overrides application instructions.

Jailbreak

A technique that manipulates an LLM into bypassing its safety alignment and producing content its guardrails are meant to block.

Prompt Injection

An attack where crafted input overrides an LLM's intended instructions, causing it to follow attacker-controlled commands instead.

Retrieval-Augmented Generation (RAG)

An architecture that grounds LLM responses in external documents retrieved at query time, improving accuracy but expanding the attack surface.

System Prompt

The developer-set instructions that define an LLM's role, rules, and behavior, sent ahead of user input in every request.