AI agent skills
PulseAugur coverage of AI agent skills — every cluster mentioning AI agent skills across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
AI agents require foundational skills in programming and software development
This cluster discusses the development of skills necessary for AI agents, drawing from recent attendance at an event. The conversation highlights the importance of foundational knowledge and good practices in creating t…
-
Developer releases local AI agent skill router for sub-20ms performance
A developer has created "Routed," an open-source, local-first hybrid router designed to improve the efficiency of AI agent skills. This tool operates entirely offline on a CPU, offering sub-20ms response times without t…
-
AI Agent Skills Ship With Security Flaws, Report Finds
A recent analysis revealed that a significant portion of AI agent skills contain security vulnerabilities. Out of nearly 4,000 evaluated skills, over a third were found to have flaws, with a notable percentage containin…
-
AI agent skills show extreme install concentration, with most being free
A recent analysis of 58,751 AI agent skills revealed a highly concentrated distribution of installs, with a Gini coefficient of 0.96. This indicates that a small fraction of skills account for the vast majority of insta…
-
AI agent skill security scanners show low agreement on safety
A study of five AI agent skill security scanners found that they agree on safety assessments less than 36% of the time. These scanners, which evaluate different security aspects like code vulnerabilities and prompt inje…
-
New benchmark reveals AI agent skills vulnerable to novel attacks
Researchers have developed SkillHarm, a new benchmark for evaluating security vulnerabilities in AI agent skills. The benchmark includes two attack scenarios: Fixed-Payload Poisoning, where a skill directly compromises …
-
AI Agent Skills Show 26% Vulnerability Rate, Lintai Tool Offers Solution
A recent paper analyzing 31,132 AI agent skills revealed that over a quarter of them contained vulnerabilities such as prompt injection or data exfiltration. This finding has led to a new practice of manually reviewing …