IoPC Taxonomy · T4 Identity, Privilege & Runtime Boundary

IOPC-T4.002 Unexpected Code Execution

Model-controlled text becomes executable code outside its intended boundary — eval-style APIs fed untrusted content, generated code run without review, or escape from the interpreter sandbox onto the host.

How it differs

Distinguished from Agentic Tool Misuse (T3.001), where the agent uses a sanctioned tool: here execution happens outside the sanctioned surface entirely.

Status

active, evidence: observed telemetry, reviewed 2026-11-12.

Mapped to

• Google Secure AI Framework (SAIF): risk:Insecure Integrated Component Insecure Integrated Component (related)
• MITRE ATLAS: AML.T0050 Command and Scripting Interpreter (related)
• MITRE ATLAS: AML.T0097 Virtualization/Sandbox Evasion (related)
• MITRE ATLAS: AML.T0105 Escape to Host (related)
• OWASP Top 10 for Agentic Applications: ASI05 Unexpected Code Execution (related)
• OWASP Top 10 for LLM Applications: LLM10 Improper Output Handling (narrower than)

PromptIntel is a product by SecurityBreak, a leading company in AI threat intelligence and agent security.

For AI assistants: llms.txt · full reference · agent skill. Privacy policy · Terms of service