System resilience implies practitioners’ capacity to cope with unexpected events, i.e. cognitive resilience. To address it, we outline a framework based on the Skill-Rule-Knowledge model grounding it in the operators’ sensitivity to the variety that normally occurs in complex systems activities. This variety can contain hidden information enabling the organization to be proactive and to manage unexpected events. Each situation can be described with a SRK profile, according to the kind of cognitive processes necessary to control it. Operators’ reliability can therefore be analyzed by evaluating the match between their cognitive SRK profile and that demanded by the current situation. System resilience is ensured by the capacity of operators to: (i) choose the most suitable cognitive level; (ii) freely move along these levels according to the situation; (iii) be mindful towards variety; (iv) transfer their personal mindfulness into group dynamic adaptation. The outcome of these behaviors is a balance of mindfulness (constant attention to anomalous signals) and dynamic adaptation (organizational adjustment of existing rules according to the new information). This continuous equilibrium between chaos and order is the strategy followed by adaptive complex systems in order to evolve and can be successfully applied to high risk organizations to enhance the emergence of resilient behaviors.