The Cognitive Crucible
A transparent, process-driven architecture for the algorithmic age.
Prompt & Polish Workspace
Instead of a blank document, students work inside a version-controlled dual-pane editor. Every interaction with the AI is tracked as a "commit." We evaluate how students direct the AI, aggressively edit drafts, and refine logic—treating algorithmic output as raw material, not a final product.
Explore the Crucible Architecture →The Metacognition Dashboard
We abandon punitive "AI Plagiarism Scores." Our dashboard provides educators with a Cognitive Engagement Score (CES) based on structural prompt complexity, verification persistence, and the algorithmic sculpting ratio (Diff Analysis).
View Dashboard Metrics →The Red Team Protocol
To combat automation bias, our autonomous Socratic agents occasionally and safely inject deliberate logical fallacies. Students must act as adversarial auditors, utilizing the "Flag" mechanic to correct the AI's flawed reasoning, proving epistemic vigilance.
Audit the Protocol →