In this paper, we show how a domain dependent know-how textual database of advices and warnings can be constructed from procedural texts. We show how arguments of type warnings and advices can be annotated and extracted from procedural texts, and propose a format and a strategy to automatically generate a know-how textual database. Categories and Subject Descriptors H.4 [Information Systems Applications]: Miscellaneous General Terms Human factors, Experimentation Keywords structure and content analysis, text semantics, automatically generated document