Prompt Security threat coverage: OWASP LLM Top 10 & MITRE ATLAS (2026)
Acquired · SentinelOnePrompt Security is a Guardrails / LLM Firewall vendor based in New York, USA. Its flagship product is Prompt Security GenAI Protection Platform (employee AI / shadow AI, homegrown app protection). Against the OWASP LLM Top 10 (2025) we score it Covered on 3 of 10 risks, Partial on 4, with 6 MITRE ATLAS techniques mapped as of . It now operates under SentinelOne.
- 3/10
- Risks covered
- 4
- Partial
- 6
- ATLAS techniques
- ~$23M total ($5M seed + $18M Series A)
- Funding
- Acquired
- Status
At a glance
- Category
- Guardrails / LLM Firewall
- Headquarters
- New York, USA
- Founded
- 2023
- Employees
- 11-50 (approx)
- Flagship product
- Prompt Security GenAI Protection Platform (employee AI / shadow AI, homegrown app protection)
- Deployment
- SaaS + on-prem/self-hosted; inline proxy, browser extension, SDK
- Best for
- Governing employee GenAI/shadow-AI usage and wrapping homegrown LLM apps with real-time DLP and injection defense
- Funding to date
- ~$23M total ($5M seed + $18M Series A)
- Last round
- Series A ($18M) · 2024
- Status
- Acquired by SentinelOne
Facts as of .
What does Prompt Security do?
A GenAI firewall pointed both ways — at employees' shadow-AI use and at your own LLM apps; acquired by SentinelOne. It ships as SaaS + on-prem/self-hosted; inline proxy, browser extension, SDK, and it's built for governing employee GenAI/shadow-AI usage and wrapping homegrown LLM apps with real-time DLP and injection defense. It's a weaker fit for model artifact scanning or ML training-pipeline supply-chain security.
Which OWASP LLM Top 10 risks does Prompt Security cover?
Prompt Security's strongest verdict is LLM01Prompt Injection. The two tables below come from our editorial read of Prompt Security's public documentation as of (it's what the docs support, not a hands-on lab test), and this row also sits in the full matrix across all 21 vendors.
| OWASP LLM Risk | What it is | Prompt Security coverage | How it's addressed | Source |
|---|---|---|---|---|
| LLM01 Prompt Injection | User or hidden input overrides the model's rules or intended behavior. | Covered | Secures homegrown apps against prompt injection in real time. | https://prompt.security/solutions/homegrown-genai-apps |
| LLM02 Sensitive Information Disclosure | The model leaks secrets, personal data, or proprietary content in its output. | Covered | Filters/obfuscates sensitive data to prevent exfiltration to 3rd-party LLMs and vector DBs. | https://prompt.security/solutions/homegrown-genai-apps |
| LLM03 Supply Chain | Compromised models, datasets, plugins, or dependencies add risk before runtime. | Not covered | No mention across product docs (homepage, product/solution pages, vulnerability taxonomy list covering Prompt Injection, Jailbreak, Privilege Escalation, Data Privacy, Toxic Content, Denial of Wallet/Service, Prompt Leak) — model/software supply-chain security is absent. | https://prompt.security/vulnerabilities/jailbreak |
| LLM04 Data and Model Poisoning | Tampered training or fine-tuning data corrupts how the model behaves. | Not covered | No mention of data/model poisoning as a product capability across homepage or vulnerability pages checked; blog has general research posts on the topic but no documented product feature. | https://prompt.security |
| LLM05 Improper Output Handling | Downstream systems trust model output without checking it, enabling injection or code execution. | Partial | Blocks RCE and manipulated outputs; response redaction/blocking. | https://prompt.security/solutions/homegrown-genai-apps |
| LLM06 Excessive Agency | An agent holds more permissions, tools, or autonomy than the task needs. | Partial | Policy enforcement over AI tool actions; agentic controls noted post-acquisition. | https://prompt.security/solutions/homegrown-genai-apps |
| LLM07 System Prompt Leakage | The system prompt or the secrets inside it get exposed to users. | Partial | Detects jailbreaks/injection aimed at revealing system context. | https://prompt.security/vulnerabilities/jailbreak |
| LLM08 Vector and Embedding Weaknesses | Flaws in RAG stores let attackers poison, extract, or infer data. | Partial | Redacts sensitive data before it reaches vector databases. | https://prompt.security/solutions/homegrown-genai-apps |
| LLM09 Misinformation | The model produces false or fabricated content that users act on. | Not covered | Hallucination/misinformation detection is absent from the documented vulnerability taxonomy and homepage feature list; platform focuses on input/output security and governance, not output factuality. | https://prompt.security/vulnerabilities/jailbreak |
| LLM10 Unbounded Consumption | Uncontrolled requests drive cost, denial of service, or model extraction. | Covered | Protects homegrown apps against Denial of Wallet / unbounded consumption. | https://prompt.security/solutions/homegrown-genai-apps |
| ATLAS Technique (ID) | Tactic | Prompt Security coverage | Notes | Source |
|---|---|---|---|---|
| LLM Prompt Injection (AML.T0051) | Initial Access / Execution | Covered | Core defense for homegrown apps. | https://prompt.security/solutions/homegrown-genai-apps |
| LLM Jailbreak (AML.T0054) | Defense Evasion | Covered | Dedicated jailbreak detection. | https://prompt.security/vulnerabilities/jailbreak |
| LLM Data Leakage (AML.T0057) | Exfiltration | Covered | Real-time DLP redaction of sensitive data to/from LLMs. | https://prompt.security/solutions/homegrown-genai-apps |
| Cost Harvesting (AML.T0034) | Impact | Partial | Denial-of-Wallet protection limits resource abuse. | https://prompt.security/solutions/homegrown-genai-apps |
| LLM Plugin Compromise (AML.T0053) | Execution | Partial | Blocks RCE and unsafe tool/plugin actions. | https://prompt.security/solutions/homegrown-genai-apps |
| Extract LLM System Prompt (AML.T0056) | Discovery | Partial | Detects extraction/jailbreak attempts targeting system context. | https://prompt.security/vulnerabilities/jailbreak |
Is Prompt Security independent, and how is it funded?
Prompt Security was acquired by SentinelOne for ~$250M reported; SEC filing lists ~$133.6M cash + 1,555,099 Class A shares + assumed options (2025-08-05; completed 2025-09-05) and now operates as a brand of SentinelOne. Packaging, pricing, and support can shift under new ownership, so confirm current terms with SentinelOne before you buy.
Prompt Security alternatives
The closest alternatives we track in Guardrails / LLM Firewall are Enkrypt AI, Lakera, Lasso Security. On the open-source side, Guardrails covers similar ground.
- Prompt Security vs Lasso Security
- Prompt Security vs WitnessAI
- Prompt Security vs Pangea
- Prompt Security vs NeuralTrust
Frequently asked questions
What is Prompt Security used for?
Prompt Security is a Guardrails / LLM Firewall vendor. Its flagship product is Prompt Security GenAI Protection Platform (employee AI / shadow AI, homegrown app protection). It ships as SaaS + on-prem/self-hosted; inline proxy, browser extension, SDK, and it's built for governing employee GenAI/shadow-AI usage and wrapping homegrown LLM apps with real-time DLP and injection defense.
Is Prompt Security independent or acquired?
Prompt Security was acquired by SentinelOne and now operates as a subsidiary, as of 2026-07-13.
How many OWASP LLM Top 10 risks does Prompt Security cover?
We score Prompt Security as Covered on 3 of the ten OWASP LLM Top 10 (2025) risks, with 4 more Partial, as of 2026-07-13. The table on this page breaks down every risk, including the ones marked Not covered or Unverified.
How is Prompt Security deployed?
Prompt Security ships as SaaS + on-prem/self-hosted; inline proxy, browser extension, SDK.
What are the best alternatives to Prompt Security?
The closest Guardrails / LLM Firewall alternatives llmthreat tracks are Enkrypt AI, Lakera, Lasso Security. On the open-source side, Guardrails covers similar ground.