OpenAI · March 11, 2026

Designing AI agents to resist prompt injection

Why it matters

OpenAI treats prompt injection as contextual social engineering and uses source-sink analysis to connect attacker-controlled content with dangerous actions. The design approach combines model resistance with deterministic limits on data transmission, navigation, tool use, sandbox communication, and user confirmation.

My takeaway: Threat-model each agent workflow as a source-to-sink graph: mark every place untrusted content enters, every sensitive value the agent can read, and every external action or transmission it can trigger. Remove unnecessary paths, scope permissions, mediate dangerous sinks deterministically, show users the exact data and action at confirmation time, and test multi-step social-engineering cases.
Keep exploring

More curated notes connected through Prompt Injection and Agent Security.

OWASP GenAI Security Project · guide

OWASP Top 10 for Agentic Applications for 2026

OWASP's community guide organizes agentic-system risk into ten categories, including goal hijacking, tool misuse, identity and privilege abuse, memory poisoning, insecure inter-agent communication, cascading failures, and rogue-agent behavior. It provides a shared taxonomy and mitigation starting point rather than a certification checklist or evidence that a deployed system is secure.

OWASP GenAI Security Project · tool

FinBot CTF Is Live: A Hands-On Companion to the OWASP GenAI Security Project

OWASP FinBot is a hands-on agentic-security CTF built around a simulated multi-agent financial-services platform with real tool access. Its challenges cover prompt injection, tool misuse, policy bypass, data exfiltration, privilege escalation, remote code execution, shared context, and compromised MCP servers.

NVIDIA AI Red Team · analysis

Mitigating Indirect AGENTS.md Injection Attacks in Agentic Environments

NVIDIA walks through a simulated Go dependency that detects Codex, writes a malicious AGENTS.md, redirects the coding task, and injects instructions intended to conceal the change from pull-request summaries. The post then maps the chain to dependency, configuration-integrity, monitoring, and guardrail controls.