OWASP Top 10 for LLM Applications

On this page 7

The OWASP Top 10 for LLM Applications is a community-maintained ranking of the ten most critical security risks in software that uses large language models, published by the Open Worldwide Application Security Project. In AI safety and application security practice, it serves as a shared vocabulary for threat modeling, design reviews and testing of chatbots, retrieval systems and AI agents.

The list does not describe model alignment in the abstract. It focuses on how LLM features fail inside real applications: untrusted text reaching the model, model output reaching other systems, and models receiving more permissions than they need. It is maintained by a volunteer working group under the OWASP GenAI Security Project and is revised as attack techniques and deployment patterns change.

What the OWASP Top 10 for LLM Applications covers

The list catalogs application-level risks, meaning weaknesses that appear when a language model is connected to users, data stores, tools and downstream code. Each entry has an identifier, a description, common vulnerability examples, prevention and mitigation strategies, and sample attack scenarios, which makes it usable as a checklist during design and audit.

The current edition, released for 2025, reflects lessons from production deployments of retrieval-augmented generation and agentic workflows. The following table summarizes its ten entries.

IDRiskCore concern
LLM01Prompt injectionCrafted input overrides intended instructions
LLM02Sensitive information disclosureModel reveals personal, proprietary or secret data
LLM03Supply chainCompromised models, datasets, adapters or libraries
LLM04Data and model poisoningManipulated training or fine-tuning data
LLM05Improper output handlingUnvalidated output passed to browsers, shells or databases
LLM06Excessive agencyToo many tools, permissions or autonomous actions
LLM07System prompt leakageSecrets or logic exposed through the system prompt
LLM08Vector and embedding weaknessesFlaws in retrieval stores and embedding pipelines
LLM09MisinformationConfident but false output relied upon by users
LLM10Unbounded consumptionUncontrolled cost, resource use or model extraction
Entries of the 2025 edition

How the list evolved between editions

The first version appeared in 2023, shortly after LLM-based chat interfaces became mainstream, and a minor revision followed the same year. The 2025 edition reorganized several entries because practitioners reported that some original categories overlapped, while new attack surfaces such as vector databases and system prompts had become significant.

The table below maps notable changes. Identifiers are not stable across editions, so teams should always cite the edition year together with the entry number in reports and tickets.

2023 entry2025 treatmentReason for change
Insecure output handlingRenamed improper output handling (LLM05)Broader wording for all downstream sinks
Training data poisoningExpanded to data and model poisoning (LLM04)Covers fine-tuning and embedded backdoors
Model denial of serviceFolded into unbounded consumption (LLM10)Adds cost abuse and denial of wallet
Model theftFolded into unbounded consumption (LLM10)Extraction often relies on high query volume
Insecure plugin designAbsorbed into excessive agency and othersPlugins evolved into general tool calling
OverrelianceReframed as misinformation (LLM09)Emphasis on hallucination and its effects
Selected changes from the 2023 to the 2025 edition

Two entries were new in 2025: system prompt leakage and vector and embedding weaknesses. The working group has also produced companion material on agentic AI threats, because autonomous agents chain many model calls and tools, which multiplies the impact of every item on the main list.

Mechanisms behind the most frequent risks

Most entries share one root cause: an LLM cannot reliably separate instructions from data. Every token in its context window can influence behavior, so text from a web page, an email or a retrieved document can act as a command. This is why prompt injection holds first place in both editions and why indirect prompt injection through retrieved content is treated as part of LLM01.

A second recurring mechanism is implicit trust in model output. When generated text is rendered as HTML, executed as SQL or passed to a shell without validation, classic flaws such as cross-site scripting, injection and remote code execution return through a new channel. A third mechanism is privilege concentration: an agent holding broad credentials turns a single manipulated response into real actions, which is the essence of excessive agency.

  • Untrusted input to the model drives prompt injection, poisoning and embedding attacks.
  • Untrusted output from the model drives improper output handling and misinformation.
  • Excess capability around the model drives excessive agency and unbounded consumption.
  • Secrets placed near the model drive sensitive information disclosure and system prompt leakage.

Applying the list in design, testing and governance

The list is most effective when it is mapped to concrete controls at each stage of the delivery lifecycle. Architects use it during threat modeling, testers convert entries into test cases for AI red teaming, and governance teams reference it in policies and vendor questionnaires. The decision table below links typical situations to recommended actions.

SituationRelevant entriesRecommended control
Chatbot reads external web pagesLLM01, LLM05Isolate retrieved text and sanitize all output
Agent can send email or pay invoicesLLM06, LLM01Least privilege and human approval for actions
RAG over internal documentsLLM08, LLM02Enforce document-level access control at retrieval
Third-party model or adapter adoptedLLM03, LLM04Verify provenance, signatures and model cards
Public API with metered model callsLLM10Rate limits, quotas and spending alerts
Answers used for medical or legal adviceLLM09Grounding, citations to source data and human review
Using the list across the lifecycle

Practitioners often combine the list with AI guardrails, content filters and conventional controls such as input validation, output encoding, secret management and logging. No single control closes an entry completely, so defense in depth remains the guiding principle.

Limitations and common misreadings

The ranking reflects expert consensus and practitioner surveys, not a statistically measured frequency of incidents. Its order indicates perceived importance, and a lower position does not mean a risk is harmless in a particular system. The list is also not a compliance standard: passing a checklist review does not certify that an application is secure.

MisreadingWhy it is wrongBetter practice
A good system prompt prevents injectionInstructions in the prompt can be overriddenTreat model output as untrusted regardless
The list covers model alignmentIt targets application security risksPair it with alignment and safety evaluations
Guardrail products solve all tenFilters miss novel phrasing and encodingsLayer architectural controls with filters
Entry numbers are permanentNumbering changed between editionsCite edition year and entry name
Frequent misunderstandings

Finally, the list deliberately stays technology-neutral. It does not prescribe specific products or model providers, so each organization must translate the guidance into controls that fit its architecture, data sensitivity and regulatory obligations.

A non-technical view

Think of the OWASP Top 10 for LLM Applications as a list of the ten most common ways AI-powered software gets into trouble. Security experts from many companies compared their experiences and agreed on which problems matter most.

Many of these problems come from one simple fact: an AI model treats everything it reads as potential instructions. A hidden sentence on a web page can therefore trick a helpful assistant into doing something its owner never intended. The list helps builders ask the right questions before launch, such as what the AI is allowed to do, what it can see, and who checks its answers.

A concrete example

A mid-sized insurance company plans to launch an AI assistant that reads customer emails, looks up policy records and drafts replies. Before release, its security team runs a two-day threat modeling workshop and walks through each entry of the 2025 edition. They discover that the assistant could read any policy in the database (LLM08 and LLM02) and send replies without review (LLM06).

The team restricts retrieval to the policies of the sender, removes the direct sending capability and adds a human approval step. During testing, a red teamer hides an instruction in an email asking the assistant to forward another customer’s records. The attack now fails, because the assistant has no access to that data and cannot act without approval. The company also sets a monthly spending cap on model calls to address unbounded consumption.