Pre-Inference-Governance für cyberkritische Systeme Action-Specific Oversight als vorgelagerte Legitimitätsarchitektur

Major cybersecurity incidents rarely result from a lack of detection or technical safeguards. Instead, they emerge from failures of decision legitimacy prior to action: unclear authorization, premature automation, responsibility gaps, and the absence of enforceable stop mechanisms.This paper introduces Pre-Inference Governance as a formal governance layer that precedes both automated and human response in cyber-critical contexts. Building on the ASiSO (Action-Specific Oversight) architecture, it defines a non-operational decision-legitimacy model structured around Pre-Inference Stops, Action Thresholds, Veto Architectures, and Accountability Ledgers.The framework does not prescribe defensive techniques, response playbooks, or system configurations. It is intentionally non-tactical and non-implementational. Its purpose is to provide a conceptual reference for evaluating when actions may legitimately occur, who is authorized to initiate them, and how accountability is preserved before irreversible impact arises.The model is domain-agnostic and does not reference specific sectors, jurisdictions, or incidents.