Common Weakness Enumeration
PulseAugur coverage of Common Weakness Enumeration — every cluster mentioning Common Weakness Enumeration across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
AI models tested on cybersecurity vulnerabilities with Gemini 4 Argon leading
Collinear AI has released version 2 of its CWE-bench, a benchmark designed to test the cybersecurity capabilities of AI models. This new version includes 100 tasks, covering 91 different Common Weakness Enumerations (CW…
-
New pipeline generates security-focused LLM code datasets
Researchers have developed RiskyPy, a new automated pipeline designed to create datasets for evaluating the security and functionality of code generated by large language models (LLMs). This system generates prompts cal…
-
New CryptoBench dataset tests LLM code review for cryptographic misuse
A new dataset called CryptoBench has been developed to evaluate the ability of large language models (LLMs) to detect cryptographic API misuse in code. The dataset consists of 54 Python snippets, including both vulnerab…
-
New VulDec Framework Uses Neural Decompilation for Binary Vulnerability Detection
Researchers have introduced VulDec, a novel framework designed to detect vulnerabilities in binary executables by employing neural decompilation. This approach addresses the challenge of analyzing code without source ac…
-
EntailLLM framework verifies LLM agent vulnerability discovery with logic programming
Researchers have developed EntailLLM, a framework designed to enhance the reliability of large language model agents in identifying software vulnerabilities, particularly in safety-critical applications like medical dev…
-
OpenRouting uses Claude LLM to triage security vulnerabilities
OpenRouting has developed a system that uses Claude, accessed via Amazon Bedrock, to triage findings from static analysis security testing (SAST) tools. Instead of relying solely on pattern matching, the LLM evaluates e…
-
New SecGent framework enhances LLM code security and functionality
A new framework called SecGent has been developed to improve the security of code generated by large language models without sacrificing functionality. This framework retrieves relevant security information from the Com…
-
New tool statically scans GGUF model files for security bugs
Munim, Inc. has released a new tool, gguf-scan.py, designed to statically identify potential security vulnerabilities in GGUF model files before they are loaded. The tool focuses on detecting integer overflow bugs in th…
-
AI-generated code shows significant security flaws in new benchmark
A new benchmark, the Last Secure Code Benchmark, has been developed to evaluate the security of code generated by AI coding agents. The benchmark, comprising 450 tasks across seven programming languages and numerous Com…
-
New Benchmark Tests AI Agents' Ability to Locate Software Vulnerabilities
Researchers have introduced the Vulnerability Localization Benchmark (VLoc Bench), a new dataset designed to evaluate the ability of AI agents to identify specific code locations associated with security vulnerabilities…
-
AI code generation security evaluated with new benchmarks and dynamic testing
Two new research papers explore the challenges of generating secure code with AI models. The first paper introduces CodeSecEval, an execution-based benchmark for evaluating secure code generation, and proposes SecAwareC…
-
New framework turns vulnerability history into executable detection rules
Researchers have developed BUGSTONE-E2E, a framework designed to transform historical vulnerability data into executable detection rules. This system mines fixing commits to create reusable rules, which are then process…
-
New framework enables targeted poisoning of AI-generated code
Researchers have developed a new framework called CodePoisonRAG that can inject specific vulnerabilities into code generated by Retrieval-Augmented Code Generation (RACG) systems. This framework works by transforming be…
-
Athena system uses knowledge graphs to identify vulnerable software libraries
Researchers have developed Athena, a novel graph-based system designed to identify libraries affected by software vulnerabilities. Unlike previous methods that treat vulnerability data as isolated text, Athena models th…
-
Anthropic integrates advanced Claude Mythos 5 into Claude Security for enterprise vulnerability scanning
Anthropic has integrated its advanced Claude Mythos 5 model into its Claude Security product, offering enterprise clients enhanced vulnerability scanning capabilities. This new feature allows teams to scan GitHub reposi…
-
New agent detects misinformation in RAG systems
Researchers have developed an "Evaluation Agent" to address the security and reliability gap in Retrieval-Augmented Generation (RAG) systems. This agent acts as middleware to detect misinformation and knowledge poisonin…
-
New CTIFoundry corpus scaffold boosts LLM agent cyber threat intelligence
Researchers have introduced CTIFoundry, a new corpus scaffold designed to enhance the capabilities of LLM agents in cyber threat intelligence (CTI). This system structures CTI data by creating an ontology graph from aut…
-
AI code generation pipeline tackles security vulnerabilities
A new research paper introduces an automated pipeline designed to detect and fix security vulnerabilities in code generated by AI development tools. The pipeline processes code from LLM-generated prompts, uses tools lik…
-
New VICBench benchmark evaluates code vulnerability detection tools · 2 sources tracked
Researchers have introduced VICBench, a new benchmark designed to evaluate code vulnerability detection tools. The benchmark comprises 100 verified vulnerability-inducing commits (VICs) across 88 projects in Python, Jav…
-
New AVE standard addresses AI agent behavioral vulnerability classification
A new open standard called AVE (Agentic Vulnerability Enumeration) has been developed to classify behavioral vulnerabilities in AI agents, addressing a gap left by existing systems like CVE and CWE. These traditional sy…