Open-source software (OSS) projects often struggle to efficiently assign issues to contributors whose skills align with task requirements. Without targeted recommendations, contributors may overlook suitable issues, leading to delayed resolutions and reduced engagement. Seeking to mitigate this barrier, previous studies proposed tagging of issues with categories of libraries as a proxy for the skills sufficient to solve them. In a different direction, researchers also proposed identifying skills from developers in open-source projects in an attempt to support managers in performing the allocation. Notwithstanding the advances, if contributors are overconfident, they still might pick an issue to solve beyond their abilities. Similarly, managers may face confirmation bias and allocate an issue incompatible with the contributor’s skill set. In addition, studies reported that maintainers have little time available to support new contributors and want contributors to have autonomy and decide about contributions. To address this, we present a fully automated, AI-powered issue recommendation system that integrates past contribution history with skill-based matching. We mine public GitHub repositories to extract contributor skills using commit histories and issue resolutions, and infer issue requirements using both traditional techniques and Large Language Models (LLMs). We evaluate three matching strategies—TF-IDF, sentence-BERT (s-BERT), and LLM-based approaches—and find that the simple TF-IDF model outperforms more complex methods, achieving a top-15 accuracy of 70%. We also explore the use of a canonical skill superset for standardizing skill representations. Our findings show that historical contribution data is a significant feature for OSS issue assignment and that lightweight lexical methods remain highly effective in specific tasks. Therefore, integrating it with other features might improve performance. This work contributes a scalable framework for personalized issue recommendation that supports diverse OSS environments and enhances contributor-task alignment.