Lack of security is a constant concern in open distributed systems. Because of this problem, many tools for evaluating vulnerabilities of networks, as well as for their protection, are being developed and largely deployed; for example, techniques for encryption, antivirus, firewall, and IDSs (Intrusion Detection Systems). Among these, there are IDSs that are increasingly conceived, designed, and implemented. Currently, IDSs are created using software agents. Although IDSs can provide intrusion detection and countermeasures against threats, they are subjected to attacks. Thus, it is necessary to ensure security for IDSs. This chapter proposes a framework for providing self-security, self-reliability of message exchange, self-reliability of components, and self-integrity for IDSs based on agents. The proposed solution is implemented as an extension of the IDS-NIDIA (Network Intrusion Detection System Based on Intelligent Agents), whose architecture has an intelligent agent society that communicates in a cooperative way in a distributed environment. Some tests show the applicability of the proposed solution.