UTC

OWASP Top 10 for LLM Applications 2026

LLM05:2026 Data and Model Poisoning

2 advisories on ShieldMap fall under it through CWE-349, and 1 technique page explains how the attacks work.

OWASP Top 10 for Agentic Applications
  • ●ASI04:2026 Agentic Supply Chain VulnerabilitiesPoisoned pre-trained models and malicious deserialization distributed via public repositories, and tampered inference-time artifacts such as chat templates/GGUF
  • ●ASI06:2026 Memory & Context PoisoningPersistent agent memory and recommendation poisoning, and long-term manipulation of agent decisions via injected memory
  • ●ASI08:2026 Cascading FailuresPoisoned inputs propagating across multi-agent and enterprise workflows into unintended data exposure, and cross-tenant contamination via shared embeddings/memory
CWE weaknesses
  • ●CWE-349 Acceptance of Extraneous Untrusted Data With Trusted DataAttacker-controlled untrusted data is mixed into trusted training corpora, retrieval stores, and continuous-learning feeds, the canonical poisoning weakness.
  • ●CWE-829 Inclusion of Functionality from Untrusted Control SphereLoading untrusted models, LoRA and PEFT adapters, tokenizer configs, and chat templates from public repositories pulls poisoned functionality into the pipeline.
  • ○CWE-494 Download of Code Without Integrity CheckSigning and hash verification of model artifacts is prescribed because unverified model downloads carry tampered weights.
  • ○CWE-1427 Improper Neutralization of Input Used for LLM PromptingPoisoned retrieval documents and hidden web instructions reach the prompt without neutralization, the indirect-injection facet of poisoning.
MITRE ATLAS tactics
  • ●AML.TA0003 Resource DevelopmentAdversaries publish poisoned datasets and models and inject tainted samples into training data that others will consume.
  • ●AML.TA0006 PersistenceBackdoors and sleeper triggers survive safety alignment and retraining, staying dormant in the model until a specific input activates them.
  • ●AML.TA0011 ImpactPoisoning erodes model and dataset integrity, degrading accuracy, weakening refusals, and steering the model toward harmful outputs.
  • ○AML.TA0004 Initial AccessOrganizations that pull compromised models or datasets from public repositories inherit the hidden triggers along with the artifact.
  • ○AML.TA0001 AI Attack AdaptationThe attacker backdoors the model and crafts trigger or adversarial data before the model is deployed.
  • ○AML.TA0005 ExecutionLoading an unsafe artifact executes attacker code, as with pickle deserialization on model load or chat-template injection in a model file.
NIST AI 600-1 risks
  • ●NIST Information IntegrityPoisoning steers answers, injects subtle misinformation, and manipulates recommendations and downstream business decisions.
  • ●NIST Value Chain and Component IntegrationThird-party datasets, shared model repositories, and bundled artifacts introduce compromise through the GenAI value chain.
  • ●NIST Information SecurityData and model poisoning, embedded backdoors, and unsafe-artifact code execution are core security risks.
  • ○NIST Dangerous, Violent, or Hateful ContentTargeted poisoning erodes refusal behaviors, as when a public chatbot was manipulated into producing offensive content.
NIST AI RMF categories
  • ○MEASURE 3 track risks over timeOutputs, training loss, and behavioral patterns are monitored for drift against defined thresholds to detect subtle poisoning that emerges over time.
  • ○MEASURE 2 evaluate trustworthinessContinuous red-teaming with adversarial and trigger-based prompts, with dedicated trigger-probing required after every alignment cycle, evaluates the model for hidden backdoors.
  • ○GOVERN 6 third-party & supply-chain policyVendor vetting, model and data lineage tracking, and controls against open-source dataset supply-chain poisoning and malicious models address third-party poisoning vectors.
OWASP GenAI Data Security
  • ●DSGAI04 Data, Model & Artifact PoisoningThe direct peer covering poisoning across pre-training, fine-tuning, embeddings, RAG, transfer learning, and distributed non-weight artifacts.
  • ○DSGAI05 Data Integrity & Validation FailuresStrict, continuous validation of training and retraining inputs is the front-line defense against injected poison.
  • ○DSGAI21 Disinformation & Integrity Attacks via Data PoisoningPoisoning a knowledge base surfaces adversary-controlled content as authoritative output, an integrity attack driven by corrupted data.
  • ○DSGAI07 Data Governance, Lifecycle & ClassificationDataset and model lineage tracked through an SBOM or ML-BOM plus version history enables rollback and forensics after a poisoning event.
OWASP AIVSS agentic risks
  • ○AIVSS-6 Agent Memory and Context ManipulationAttackers embed hidden instructions in web content that poison an agent's persistent memory or recommendations across sessions, so the agent begins prioritizing attacker-controlled logic.
  • ○AIVSS-8 Agent Supply Chain and Dependency RiskBackdoored models and unsafe-serialization artifacts distributed through public repositories carry hidden triggers that a downstream agent inherits when it loads them.
  • ○AIVSS-3 Agent Cascading FailuresPoisoned inputs propagate across multi-agent and enterprise workflows, and shared embeddings or memory layers spread contamination from one tenant to others.
CSA AI Controls Matrix domains
  • ●MDS Model Development Security
  • ●DSP Data Security and Privacy Lifecycle Management
  • ○STA Supply Chain Management, Transparency, and Accountability
  • ○LOG Logging and Monitoring
  • ○TVM Threat & Vulnerability 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.