Sign in
All library
Library· Security

Securing your agents

Prompt injection, excessive agency, and the least-privilege patterns from the OWASP LLM Top 10.

#security#ops
6 min read

Agents that read untrusted content and call tools are a new class of attack surface. Treat every tool call as if the model were a partially-trusted user.

Top risks (OWASP LLM Top 10)

  • LLM01 — Prompt injection (OWASP) — the model follows instructions hidden in retrieved documents, emails, or web pages.
  • LLM02 — Insecure output handling — using model output directly in shell, SQL, or HTML.
  • LLM06 — Excessive agency — the agent has tools with broader permissions than the task requires.

Defenses

  • Least privilege — scope tool credentials to the user and the action. Never give one agent a global admin key.
  • Human-in-the-loop for irreversible actions (payments, deletes, sends).
  • Input isolation — clearly delimit untrusted content (<untrusted>...</untrusted>) and instruct the model to never follow instructions inside.
  • Output validation — parse and validate tool arguments before executing.
  • Rate + spend limits per user and per session.

Deeper reading