OWASP
PulseAugur coverage of OWASP — every cluster mentioning OWASP across labs, papers, and developer communities, ranked by signal.
- developed by OWASP Cornucopia 90%
- instance of EchoLeak 90%
- instance of OWASP Cornucopia 70%
- used by Common Weakness Enumeration 70%
- instance of Microsoft Copilot for Microsoft 365 70%
- affiliated with Google Summer of Code 2018 70%
- developed by Bifröst 70%
- instance of SQL injection 70%
- used by Javascript 60%
- affiliated with MITRE ATLAS 60%
- affiliated with DEV Community 60%
- other EchoLeak 60%
- 2026-08-21 research_milestone OWASP is hosting a virtual conference featuring a talk on AI threat modeling. source
- 2026-08-11 research_milestone OWASP released the Top 10 for LLM Applications 2026, a guide to critical security risks for LLM-powered applications. source
- 2026-05-13 research_milestone OWASP added memory poisoning as a critical vulnerability (ASI06) to its Top 10 for Agentic Applications. source
12 day(s) with sentiment data
-
OWASP Top 10 for LLM Apps flags prompt injection as top threat
The OWASP Top 10 for LLM Applications list highlights critical security vulnerabilities in AI and large language model applications. Prompt injection is identified as the primary threat, underscoring the risks associate…
-
OWASP releases Top 10 risks for agentic AI applications
OWASP has released a "Top 10 for Agentic Applications" detailing ten risks associated with AI systems. This list, published in December 2025, aims to highlight potential security vulnerabilities in agentic AI. Separatel…
-
OWASP Agentic AI Top 10 risks explained for individual users
A developer has translated the OWASP Agentic AI Top 10 risks into practical advice for individuals running multiple agents on a personal laptop. The advice emphasizes securing agent actions at the tool-call level rather…
-
Maxim AI releases Bifrost, an open-source AI gateway for LLM jailbreak detection
An open-source AI gateway called Bifrost has been developed by Maxim AI to address the increasing threat of LLM jailbreaking. This gateway centralizes security by inspecting both user inputs and model outputs, aiming to…
-
AI Guardrails: Bifrost Leads Gateway Enforcement Amidst Evolving Threats
Enterprise AI guardrails are crucial for securing production LLM deployments against vulnerabilities like prompt injection and data exfiltration. Solutions range from in-application frameworks like NVIDIA NeMo Guardrail…
-
OWASP releases new guidelines for securing enterprise AI
OWASP has released new guidelines aimed at securing and validating artificial intelligence systems within businesses. The updated resources are intended to help organizations implement AI responsibly and address potenti…
-
Registry indexes 66K MCP servers, purges fake install counts
A new registry has been launched to track MCP servers across GitHub, npm, and PyPI, indexing over 66,000 servers and identifying over 130,000 tracked ecosystems. The registry purges inaccurate data, including 2,910 inve…
-
Agent identity and scoped credentials insufficient for AI security
A recent discussion highlights that while distinct agent identities and scoped credentials are vital for security, they do not fully address the risks in agent systems. These measures, which trace actions to named agent…
-
OWASP releases new editions of its mobile app threat modeling game
OWASP has released version 2.0 of its Cornucopia Mobile App Edition, a card game designed to aid in threat modeling for mobile applications. This new edition is compatible with the latest versions of MASVS, MASTG, and M…
-
OWASP reveals top 10 security risks for AI agent skills
The OWASP Agentic Skills Top 10 (AST10) report identifies the ten most critical security risks associated with AI agent skills. This initiative aims to highlight vulnerabilities across major AI agent platforms, includin…
-
New WAPP framework enables safe learning of WAF security policies
Researchers have developed a new framework called WAPP (Whitelisting Autonomous Policy Producer) designed to safely learn positive security policies for Web Application Firewalls (WAFs) from live traffic. Traditional WA…
-
Untracked AI servers exploit developer tools, posing security risks
Untracked Model Context Protocol (MCP) servers are spreading rapidly on developer machines, posing significant security risks. These servers, often installed via simple JSON file edits, run with inherited developer cred…
-
New tool tackles high false positive rates in MCP security scanners
A new tool called mcpaudit has been developed to address the high false positive rate in Machine Configuration Protocol (MCP) security scanners. These scanners, often relying on simple pattern matching, incorrectly flag…
-
AI Governance Gap Poses Board-Level Threat Amidst Rapid Investment
Organizations are investing heavily in AI, with global investment projected to reach $2.5 trillion, yet many are deploying these technologies without adequate governance, security, or accountability. This gap between ra…
-
AI Agents Lack Identity and Accountability, Exposing Security Risks
Organizations are increasingly deploying AI agents without robust identity and accountability frameworks, akin to early API security practices. Unlike human employees who undergo verification, role definition, and overs…
-
OWASP launches OASIS to speed up open-source vulnerability fixes with AI
The Open Web Application Security Project (OWASP) has introduced OASIS, a new initiative designed to accelerate the patching of vulnerabilities in open-source software. OASIS combines AI-generated code patches with huma…
-
RAG poisoning emerges as a supply chain attack targeting LLM knowledge bases
Researchers have identified a new type of attack called RAG poisoning, which targets the knowledge base of Retrieval-Augmented Generation (RAG) systems during the indexing phase, rather than at query time like prompt in…
-
AI agents' generate-then-execute pattern creates critical security risks
AI agents executing LLM-generated code without external validation present a significant security risk, as demonstrated by a recent incident where an agent deleted production data due to an unreviewed query. This "gener…
-
Indirect prompt injection emerges as critical LLM vulnerability
Indirect prompt injection, a vulnerability where malicious instructions are hidden within data fetched by AI agents, has been identified as a critical threat. This attack vector bypasses LLM defenses because instruction…
-
MCP protocol vulnerability allows server-supplied instructions to compromise AI agents
A significant security vulnerability has been identified in the MCP protocol, allowing malicious servers to inject harmful instructions into AI agents. This exploit, termed 'line jumping' by Trail of Bits, occurs when s…