Misinformation detectors still treat online discourse as context-free text. They excel at spotting fallacies yet misfire when power, identity, or narrative roles enter the conversation. Seven recent failures expose a systematic blind spot. Feminist technocognition is introduced as a three-level extension that links social context to algorithmic inference through context-rich mappings of cognitive-science insights, standpoint annotation grounded in feminist standpoint theory, and tailored interventions for equitable machine-learning models. The resulting architecture yields four falsifiable predictions. Practical implications are outlined for data curation, social analysis, and policy mechanisms requiring bias audits to report epistemic as well as demographic skew. Feminist technocognition thus promises misinformation defenses that are both more accurate and more equitable than current alternatives, expanding the research horizon for humanities scholars, AI-ML engineers, cognitive scientists, and digital policy makers.