Hostile Post detection on social media is a crucial task that aims to identify harmful content such as hate speech, cyberbullying, fake news and offensive language. This survey provides a detailed survey of existing and emerging techniques for this task, including machine learn-ing techniques, deep learning techniques, and hybrid techniques that integrate multiple ap-proaches to improve model performance. We also explore the use of reinforcement learning, genetic algorithms, and pseudo-labeling techniques for optimising performance. There are several challenges identified, including dataset limitations, annotation inconsistencies, model interpretability, and ethical issues such as fairness and cultural sensitivity. The review also examines future trends in multilingual hostile post detection, cross-platform adaptability, and the increasing use of transformer-based architectures. Finally, we emphasize the importance of developing scalable, explainable, and context-aware solutions, and we outline future directions for research. This work is intended to serve as a reference for scholars and practitioners working toward safer online environments.