This work introduces a simple yet effective approach to integrate Automatic Post-editing (APE) with Word Level Quality Estimation (QE) using an encoder-decoder Transformer architecture, which we term error-prompted APE. The model employs self-detected translation errors by Word Level QE as prompts to refine translations. Specifically, the error-prompted APE model is self-contained: its encoder functions as a quality estimator, detecting errors in the input translation; the decoder then leverages these identified errors as prompts to refine the translation accordingly. This detailed error prompts enable our model to selectively correct errors while preserving the integrity of the rest of the content. We conduct extensive experiments on English-German and English-Chinese datasets from the WMT shared APE task. The results demonstrate that our proposed approach achieves substantial improvements.