ejentum-unified

What problem does it solve?

It solves the problem of choosing the correct reasoning, coding, honesty-under-pressure, or memory harness for a task so the model applies the right cognitive procedure and response shape instead of guessing.

Core Features & Use Cases

• Automatic multi-mode routing: Selects the appropriate harness (reasoning, code, anti-deception, memory) based on task intent and failure risk.
• Two-mode stacking for cross-domain work: Runs a primary and optional secondary mode (max two) to cover multi-dimensional tasks without signal competition.
• Consistent cross-mode response handling: Provides a shared semantic response structure (prohibitions, control flow, topology, assertions, and amplify/suppress signals) to improve absorption and execution.

Example use case: You’re debugging a complex feature where correctness matters and user framing might hide issues; route to the code harness for engineering procedure, optionally pair with anti-deception to avoid rubber-stamping under persuasive context.

Quick Start

Tell your AI agent the task in one sentence and specify the mode as reasoning, code, anti-deception, or memory, then follow the router’s absorption steps before producing your final output.