Operator / Platform ErrorsDocumentation Index
Fetch the complete documentation index at: https://docs.svrnos.com/llms.txt
Use this file to discover all available pages before exploring further.
Definition
Adversarial input that defeated or bypassed Agent A’s input filter propagates through the agent chain to peer agents that do not re-filter, causing system-wide compromise from a single entry point.Distinct from
- GER-421 — Adversarial payload propagates through agent chain → this code. Single defeat event at one entry point → GER-421.
Anchor incidents (5)
GPT-3-Based Twitter Bot Hijacked Using Prompt Injection Attacks
secondary — AIID #352 · 2022-09-15
Bing Chat's Initial Prompts Revealed by Early Testers Through Prompt Injection
secondary — AIID #473 · 2023-02-08
Jailbroken Lovable AI Allegedly Used to Generate and Host Phishing Pages, Steal Credentials, and Bypass Security
secondary — AIID #1016 · 2025-04-09
OpenAI ChatGPT Models Reportedly Jailbroken to Provide Chemical, Biological, and Nuclear Weapons Instructions
secondary — AIID #1238 · 2025-10-10
Anthropic's Claude Was Reportedly Jailbroken To Allegedly Help Steal Sensitive Mexican Government Data
secondary — AIID #1430 · 2025-12-01
Tags
multi-agent · prompt-injection