UTC

OWASP Top 10 for LLM Applications 2026

LLM10:2026 Improper Output Handling

118 advisories on ShieldMap fall under it through CWE-116, CWE-79, and 1 technique page explains how the attacks work.

OWASP Top 10 for Agentic Applications
  • ●ASI02:2026 Tool Misuse & Exploitation2026 addition, justified against this entry's own content: a general-purpose LLM passes its response to a privileged extension without output validation, causing the extension to be misused and shut down.
  • ●ASI05:2026 Unexpected Code Execution (RCE)Foundational crosswalk mapping; matches this entry's core risk of unvalidated LLM output reaching a shell, `exec`, or `eval`, resulting in remote code execution.
  • ●ASI09:2026 Human-Agent Trust ExploitationCanonical crosswalk baseline mapping — ASI09 lists Improper Output Handling as a contributing LLM risk. The content match is loose: this entry's scenarios are machine-to-machine (unvalidated output reaching a shell, browser, or database), while ASI09 concerns deceiving a human operator; the shared thread is unvalidated model output reaching a downstream trust decision.
CWE weaknesses
  • ●CWE-1426 Improper Validation of Generative AI OutputModel output is passed downstream with insufficient validation, sanitization, or handling, the entry's exact subject.
  • ●CWE-116 Improper Encoding or Escaping of OutputMissing context-aware output encoding across HTML, JavaScript, SQL, and terminal sinks is the root cause.
  • ●CWE-79 Cross-site Scripting (XSS)Cross-site scripting is the most-repeated concrete sink, addressed by content-security-policy and output-encoding defenses.
  • ○CWE-707 Improper NeutralizationThe neutralization pillar covers control-character sanitization of ANSI, BEL, OSC, backspace, and carriage-return sequences and SQL escaping and parameterization.
  • ○CWE-200 Exposure of Sensitive Information to an Unauthorized ActorUnhandled output leaks sensitive conversation and website content to an attacker.
  • ○CWE-201 Insertion of Sensitive Information Into Sent DataSensitive data is smuggled into the hostname or query string of an auto-fetched Markdown-image request.
MITRE ATLAS tactics
  • ●AML.TA0005 ExecutionUnvalidated model output passed to a shell, browser, or SQL interpreter executes adversary-controlled code, yielding command execution, cross-site scripting, or SQL injection.
  • ○AML.TA0010 ExfiltrationModel output encoded and sent to an attacker server, or embedded in a Markdown-image URL, carries sensitive data out of the application.
  • ○AML.TA0011 ImpactUnhandled output reaching a privileged sink can delete database tables or force a downstream service offline.
  • ○AML.TA0012 Privilege EscalationWhen the model holds privileges beyond the end user, unvalidated output reaching a privileged sink escalates those privileges to the attacker.
NIST AI 600-1 risks
  • ●NIST Information SecurityUnhandled model output flows into downstream systems as remote code execution, XSS, SQL injection, SSRF, CSRF, and privilege escalation.
  • ○NIST Data PrivacyUnsanitized output can leak sensitive conversation data to an attacker.
  • ○NIST ConfabulationGenerated code can reference hallucinated nonexistent software packages, a confabulation facet of unsafe output.
  • ○NIST Value Chain and Component IntegrationThird-party extensions that fail to validate inputs and hallucinated-package installs implicate component integration and supply chain.
OWASP GenAI Data Security
  • ●DSGAI12 Unsafe Natural-Language Data Gateways (LLM-to-SQL/Graph)LLM-generated SQL executed without parameterization or scrutiny reaches the database directly and can delete entire tables.
  • ○DSGAI06 Tool, Plugin & Agent Data Exchange RisksUnvalidated model output crosses into a downstream tool or extension for execution.
  • ○DSGAI01 Sensitive Data LeakageUnescaped output rendered by a client encodes and sends conversation data to an attacker-controlled server, as in image-URL exfiltration.
  • ○DSGAI16 Endpoint & Browser Assistant OverreachClient renderers such as chat UIs, IDEs, and email clients auto-fetch external resources from raw model output, executing the exfiltration.
OWASP AIVSS agentic risks
  • ○AIVSS-7 Insecure Agent Critical Systems InteractionUnvalidated model output reaching a shell, exec/eval, SQL, or file-path sink yields remote code execution, SQL injection, or path traversal on backend systems.
  • ○AIVSS-1 Agentic AI Tool MisuseA general-purpose LLM's unvalidated response passed to a privileged extension causes the extension to be misused, for example forced into an unintended shutdown.
CSA AI Controls Matrix domains
  • ●AIS Application & Interface Security
  • ○LOG Logging and Monitoring
  • ○IAM Identity & Access Management

Mappings and their reasons from the OWASP Top 10 for LLM Applications 2026, Appendix A (OWASP GenAI Security Project), shared under CC BY-SA 4.0. Mappings restructured as data; the ATT&CK section and the AICM relevance text left out. Navigation, not a compliance claim.