xby-structured-argumentation

Large language models often default to a single linear narrative when faced with complex questions, skipping the systematic deconstruction of competing positions and the dialectical testing that rigorous reasoning demand. xby-structured-argumentation fills that gap by plugging into DeepSeek Harness as an MCP server and making the internal structure of argumentation explicit: thesis, antithesis, and synthesis become first-class objects the model can create, critique, and stitch together. The structuredArgumentation tool lets you decompose any proposition into claim, premises, and conclusion; label each node as a thesis, antithesis, synthesis, objection, or rebuttal; and wire up logical links through respondsTo, supports, and contradicts fields. Confidence scores, strengths, and weaknesses attached to every node make it easy to audit how solid each step of the dialectic really is, and the nextArgumentNeeded flag helps the model know when the exchange is not yet complete. A complementary set_xby_apikey tool handles persistent API-key configuration right inside the chat session. The plugin is a natural fit for anyone who needs the model to reason rather than just recite: ethical dilemmas with no clean answer, policy proposals touching multiple stakeholders, scientific hypotheses that invite counter-examples, or any workflow where the deliverable is an inspectable chain of reasoning instead of a single verdict.