As I see it, these is no theory of social cognition, but there many theories that elaborate on social cognition.
Bandura wrote his social learning theory in 1977 and his socio cognitive theory in1986. Bandura elaborates on the concept of self-efficacy mainly in his latter book.
Social Cognitive Theory isn't to be confused with social cognition. The former is a theory of social learning proposed by A. Bandura, the latter is a broader subfield of psychology interested in how cognitive processes shape and are shaped by our social interactions.
As for Maha's question, I believe that self-efficacy is a component of Social Cognitive Theory. As I understand it, according to this theory, we learn portions of our knowledge through the observation of models (e.g. the bobo doll experiment). The theory has 4 components: Modeling, identification, self-efficacy, and outcome expectancy. In short, we observe the behavior of models (modeling) with which we can identify (identification), and, given that we believe we are capable of imitating this behavior (self-efficacy), we expect that we would obtain similar outcomes (outcome expectancy). I am not a expert in social cognitive theory, but you can access some of his work first hand through this website: https://www.uky.edu/~eushe2/Bandura/BanduraPubs.html
After looking more into the subject, I found that self-efficacy is an independent theory, which's course an important concept in the theory of Social cognitive. But The Social cognitive theory is more general. This consistent with what colleagues have referred to here during this debate. Thanks so very much for your response.