DeepTeam is a framework to red team LLMs and AI agents.
.
A-Vamshi4ad8f41ad8e7
diff --git a/docs/docs/frameworks-introduction.mdx b/docs/docs/frameworks-introduction.mdxindex c254735e..68ecd275 100644--- a/docs/docs/frameworks-introduction.mdx+++ b/docs/docs/frameworks-introduction.mdx@@ -13,35 +13,32 @@ DeepTeam supports multiple frameworks — from dataset-based testing (_BeaverTaiHere are the list of frameworks available in `deepteam`:-- [MITRE ATLAS](#mitre-atlas)-- [NIST AI RMF](#nist-ai-rmf)- [OWASP Top 10 for LLMs](#owasp-top-10-for-llms)+- [NIST AI RMF](#nist-ai-rmf)+- [MITRE ATLAS](#mitre-atlas)- [BeaverTails](#beavertails)- [Aegis](#aegis)-### MITRE ATLAS+### OWASP Top 10 for LLMs-The MITRE ATLAS framework integrates the MITRE ATLAS knowledge base, focusing on adversarial tactics and techniques used against AI systems.-It evaluates system resilience across attack phases like Reconnaissance, Resource Development, Initial Access, and Impact.+The OWASP Top 10 for LLMs framework identifies the most critical security risks in LLM applications.+It reflects the 2025 OWASP edition, covering vulnerabilities in RAG systems, agents, and model integrations.-- Tests adversarial behavior patterns from the ATLAS taxonomy-- Detects vulnerabilities such as prompt injection, data poisoning, and exfiltration-- Ideal for AI security simulation and penetration testing+- Tests for prompt injection, system prompt leakage, vector weaknesses, and more+- Simulates real-world exploit attempts and harmful outputs+- Ideal for application-level AI security assessments```python-from deepteam.frameworks import MITREATLAS-from deepteam import red_team-from somewhere import your_model_callback--atlas = MITREATLAS(num_attacks=10)+from deepteam.frameworks import OWASPTop10+owasp = OWASPTop10(num_attacks=10)risk = red_team(model_callback=your_model_callback,- framework=atlas+ framework=owasp)```-[Learn more about MITRE ATLAS](/docs/red-teaming-frameworks-mitre)+[Learn more about OWASP Top 10 for LLMs](/docs/frameworks-owasp-top-10-for-llms)### NIST AI RMF@@ -61,28 +58,31 @@ risk = red_team()```-[Learn more about NIST AI RMF](/docs/red-teaming-frameworks-nist)+[Learn more about NIST AI RMF](/docs/frameworks-nist-ai-rmf)-### OWASP Top 10 for LLMs+### MITRE ATLAS-The OWASP Top 10 for LLMs framework identifies the most critical security risks in LLM applications.-It reflects the 2025 OWASP edition, covering vulnerabilities in RAG systems, agents, and model integrations.+The MITRE ATLAS framework integrates the MITRE ATLAS knowledge base, focusing on adversarial tactics and techniques used against AI systems.+It evaluates system resilience across attack phases like Reconnaissance, Resource Development, Initial Access, and Impact.-- Tests for prompt injection, system prompt leakage, vector weaknesses, and more-- Simulates real-world exploit attempts and harmful outputs-- Ideal for application-level AI security assessments+- Tests adversarial behavior patterns from the ATLAS taxonomy+- Detects vulnerabilities such as prompt injection, data poisoning, and exfiltration+- Ideal for AI security simulation and penetration testing```python-from deepteam.frameworks import OWASPTop10-owasp = OWASPTop10(num_attacks=10)+from deepteam.frameworks import MITREATLAS+from deepteam import red_team+from somewhere import your_model_callback++atlas = MITREATLAS(num_attacks=10)risk = red_team(model_callback=your_model_callback,- framework=owasp+ framework=atlas)```-[Learn more about OWASP Top 10 for LLMs](/docs/red-teaming-frameworks-owasp-top-10-for-llms)+[Learn more about MITRE ATLAS](/docs/frameworks-mitre-atlas)### BeaverTails@@ -102,7 +102,7 @@ risk = red_team()```-[Learn more about BeaverTails](/docs/red-teaming-frameworks-beavertails)+[Learn more about BeaverTails](/docs/frameworks-beavertails)### Aegis@@ -122,7 +122,7 @@ risk = red_team()```-[Learn more about Aegis](/docs/red-teaming-frameworks-aegis)+[Learn more about Aegis](/docs/frameworks-aegis):::tipYou can customize and add more attacks and vulnerabilities to already existing frameworks to specialise red-teaming to your LLM's use case.diff --git a/docs/docs/frameworks-mitre-atlas.mdx b/docs/docs/frameworks-mitre-atlas.mdxindex 787be947..8cc3caf9 100644--- a/docs/docs/frameworks-mitre-atlas.mdx+++ b/docs/docs/frameworks-mitre-atlas.mdx@@ -125,8 +125,6 @@ Reconnaissance testing ensures your model does not reveal **sensitive internal l- `Leetspeak`, `ROT13`- `PromptInjection`, `CrescendoJailbreaking`-**Example (run only resource-development tests):**-```pythonfrom deepteam.frameworks import MITREfrom deepteam import red_team@@ -324,5 +322,3 @@ This stage ensures your system's final outputs and decisions remain aligned with- [MITRE ATLAS™ Official Documentation](https://atlas.mitre.org/)- [MITRE ATT&CK® Framework](https://attack.mitre.org/)- [NIST AI RMF (for complementary evaluation)](https://www.nist.gov/itl/ai-risk-management-framework)--Would you like me to include a **mapping table** at the end (like NIST's measure summary) showing how each MITRE tactic maps to DeepTeam vulnerabilities and attacks (a compact summary table)? It would make it ready for production docs.diff --git a/docs/docs/frameworks-nist.mdx b/docs/docs/frameworks-nist-ai-rmf.mdxsimilarity index 99%rename from docs/docs/frameworks-nist.mdxrename to docs/docs/frameworks-nist-ai-rmf.mdxindex 975e876c..a4d5a5c1 100644--- a/docs/docs/frameworks-nist.mdx+++ b/docs/docs/frameworks-nist-ai-rmf.mdx@@ -1,5 +1,5 @@----id: frameworks-nist+id: frameworks-nist-ai-rmftitle: NIST AI RMFsidebar_label: NIST AI RMF---diff --git a/docs/sidebars.js b/docs/sidebars.jsindex e1f13e0f..d022085d 100644--- a/docs/sidebars.js+++ b/docs/sidebars.js@@ -182,7 +182,7 @@ module.exports = {items: ["frameworks-introduction","frameworks-owasp-top-10-for-llms",- "frameworks-nist",+ "frameworks-nist-ai-rmf","frameworks-mitre-atlas","frameworks-beavertails","frameworks-aegis",