← Real-world cases

Zscaler ThreatLabz — web indirect prompt injection targeting AI agents in the wild

Real-world incident02 Jul 2026

Zscaler ThreatLabz (Ashwathi Sasi, Kartik Dixit, Akshay Kumar Adimulam) reported two real-world campaigns that weaponise indirect prompt injection against AI agents that browse the web, rather than lab proof-of-concepts. In both, attackers used SEO poisoning to rank a malicious page highly, then buried prompt-style instructions where a human never sees them — text moved off-screen with CSS (e.g. left:-9999px), hidden div elements, and structured JSON-LD schema metadata that machines read as trusted context. Campaign 1 impersonated documentation for a Python library ('requests-secure-v2') and instructed any agent working a coding task that it must buy a $3 API license key to fix an error, walking it through paying an attacker's Stripe or cryptocurrency wallet for a non-existent key. Campaign 2 registered a DeBank impersonation domain (reported as debank[.]auction) that used keyword stuffing and false metadata plus embedded injections telling agents to treat it as the authoritative DeBank and rank it first. ThreatLabz ran its own autonomous agent against the sites across 26 large language models: 4 were manipulated into executing the fraudulent payment (including versions of Meta's Llama and Google's Gemini) and, for the second campaign, 2 models miscategorised/endorsed the fraudulent site when they lacked reference information. Per ThreatLabz, the Ethereum wallet tied to Campaign 1 had reportedly received payments, though noted as for larger amounts than the scam's advertised key, so confirmed losses directly attributable to this scam are not established. This is among the first publicly documented in-the-wild (not PoC) monetised web indirect-prompt-injection operations aimed at general-purpose browsing and coding agents. Payload details are illustrative, not operational.

Practise the risk class — related scenarios

Interactive simulations of the risk class this case illustrates (not a re-enactment of this specific event).

🌀The Refund That Never Existed

A support chatbot invents a policy — and the company is held to it

🔑The Agent With the Master Key

An ops agent gets one god-mode credential — and one misread wipes production

🪄The Approval That Lied

A coding agent asks to write ./notes.txt — the file it actually overwrites is your SSH keys

📧The Email That Gave Orders

A support email hides instructions — and the assistant obeys them

🗄️When the Query Bites Back

A text-to-SQL agent runs the model's output straight at the database

🕵️Lies in the Loop

A poisoned issue makes the agent lie to the human who approves its actions

🪤The Bug Report That Ran Code

A fake Sentry error report hijacks a developer's coding agent into running a shell command

📼The Compromised Flight Recorder

The forensic record is itself the attack surface — an agent's log is poisoned, then quietly rewritten

👻The Email That Rewrote Its Memory

A newsletter the user asked to summarise quietly writes a false 'fact' into the agent's long-term memory — and it detonates weeks later

👁️The Invisible Webpage Command

A shopping page tells the agent to do something the user never asked for

🕵️The Logs That Lied

An attacker plants prompt injection in the audit trail — so the LLM that hunts them erases the evidence

🧠The Memory That Wouldn't Die

A single poisoned document plants a standing instruction that survives every reset

🖼️The Picture That Whispered

A screenshot that's harmless at full size becomes an order once the system shrinks it

🛡️The Watcher Watched

The eval gate that was supposed to catch the agent is itself the thing being attacked

🪪The Worker Who Spoke for the Boss

A poisoned web page hijacks a research agent — and the planner acts on its behalf

🖱️What You Click Is Not What You Get

A GUI agent clicks 'Continue' — but the screen moved, and it lands on 'Send'

🖼️Zero-Click Leak by Picture

An inbox summary quietly ships a secret to an attacker's server

More cases on Indirect Prompt Injection

EchoLeak — Microsoft 365 Copilot zero-click (CVE-2025-32711)Indirect prompt injection coined (Greshake et al.)Agentic-browser indirect-injection demos (ChatGPT Operator)ChatGPT persistent-memory exfiltration (Rehberger / 'SpAIware')MCP tool-poisoning PoC (Invariant Labs)Taxonomy of Failure Modes in Agentic AI Systems (Microsoft)ForcedLeak — Salesforce Agentforce CRM exfiltration (CVSS 9.4, no CVE)ServiceNow Now Assist — second-order prompt injection via agent-to-agent discoveryShadowLeak — ChatGPT Deep Research zero-click service-side exfiltrationIDEsaster — AI coding IDEs/agents turned into exfiltration & RCE surfacesGitHub Copilot / VS Code RCE via prompt injection ('YOLO mode', CVE-2025-53773)Agent-in-the-Middle — abusing A2A agent cards (Trustwave SpiderLabs)Agent Session Smuggling in A2A systems (Unit 42)Morris II — zero-click self-replicating adversarial-prompt worm across GenAI agentsAnamorpher — image-scaling prompt injection against production AI systemsThe Attacker Moves Second — adaptive attacks bypass 12 jailbreak/injection defenses (Nasr, Carlini et al.)MCPTox: tool-poisoning benchmark over real-world MCP serversAgentjacking — hijacking AI coding agents via Sentry error reports (Tenet Security)SearchLeak — Microsoft 365 Copilot one-click data theft (CVE-2026-42824)ChatGPhish — ChatGPT web-summary rendering turned into a phishing surfacePoisoning Claude Code: one GitHub issue hijacks the claude-code-action CI supply chainCursor 'DuneSlide' — indirect prompt injection escapes the IDE sandbox to zero-click RCE (CVE-2026-50548 / CVE-2026-50549)Context Contamination: passive prompt injection poisons LLM security-log analysisAzure DevOps MCP confused-deputy — hidden PR comments hijack AI review agentsMemGhost: stealthy one-email memory injection in persistent personal agentsAgentic botnets via universal, transferable adversarial HalluSquattingGPT-Red self-play red-teaming and the 'Fake Chain-of-Thought' injection class

AI RiskAtlas is an educational model of how GenAI & agentic systems work and fail. Architectures and payloads are illustrative and simplified for learning — not operational guidance. Real-world cases are summarised from public reporting.

Sources & further reading →·Built by Shi Yuan ↗