» Tag
research
57 postsSecurity Flaws in AI Agent Frameworks Expose Deeper Issues
Check Point reveals critical security flaws in AI frameworks, stressing the need for systemic fixes.
PromptFiction Flaw: Auto-Submitted Hidden Prompts in Claude Desktop
Oasis Security's research uncovered the PromptFiction flaw in Claude Desktop, enabling commands to be executed without user approval.
Exploiting LLM Vulnerabilities: Accessing Dangerous Instructions
Dave Kuszmar revealed vulnerabilities in LLMs that allowed access to dangerous instructions, indicating a critical security issue across the industry.
CommitBrief — AI code reviews, right in your terminal
A provider-agnostic, local-first CLI that reviews your staged changes, a historic range, or a whole GitHub pull request. Zero telemetry, no server. Free and open source.
commitbrief.comAI Science Workbench Requires a Reproducibility Graph, Not Just Chat History
Anthropic's Claude Science offers a new approach to scientific research with enhanced auditability.
Solving 20 Erdős Problems with 20 Codex Accounts Running in Parallel
A significant advancement in solving Erdős problems; a new approach for Problem 123 was developed.
New Brands Cited 0% in AI Responses: Insights from 500 Queries
New brands are cited 0% in AI responses. Wikipedia presence is crucial for visibility.
Four Eras of Cloud Security: Tools and Evaluation
Scott Piper's retrospective on four eras of cloud security research provides key insights for engineers.
Can AI Conduct Novel Security Research? Introducing the HTTP Terminator
The HTTP Terminator project explores AI's ability to develop new attack techniques, offering significant insights for security engineers.
Copilot reveals how to hack itself through manipulation
Microsoft Copilot was manipulated by researchers to reveal its own vulnerabilities, named "CoSnitch" by Varonis.
LLM-Generated Tests Influence Implementation Selection Outcomes
Key findings on the impact of LLM-generated tests on implementation selection and evaluation.