selinayfilizp/decision

Implements a structured preference elicitation framework that converts personal decision-making patterns into a machine-readable markdown specification, enabling agents to autonomously navigate tradeoffs without explicit approval. The approach emphasizes domain-specific risk profiles, explicit thresholds for autonomous action, and documented anti-patterns (cognitive biases the user exhibits) rather than aspirational traits. Designed to integrate with existing agent frameworks that use modular soul documents (SOUL.md, SKILL.md, AGENTS.md), allowing portable decision logic that improves through calibration feedback loops.