Cognitive Library & Recipes

Complex analytical execution requires forced procedural steps, not narrative suggestions.

The fundamental gap between standard LLM interfaces and the Brainstorm Reactor lies in the strict enforcement of procedural logic. Generic generative models fail complex tasks because they blur ideation, criticism, and structuring into a single massive context window. The Brainstorm Reactor solves this by enforcing an architectural division of labor via Methods, Recipes, and Workflow Vessels.

1. The Method Library

In the operational backend, the Method Library operates as a locked repository of more than 315 mathematically optimized algorithms detailing exactly how to approach a problem.

When a user initiates an exploration, they are not relying on the model’s internal stylistic preference. The Orchestrator natively injects a heavily constrained <cognitive_framework> payload directly into the system architecture before the user's intent is read.

For instance, invoking a Socratic Method protocol does not merely prompt the engine to "act Socratic." It explicitly bounds the execution into rigid operational phases:

1. Identify core assertions embedded in the input.
2. Formulate 3 distinct contradictions exposing mathematical or logical paradoxes.
3. Present only the contradictions as an actionable hierarchy.

Methods are categorized across definitive clusters representing the underlying problem morphology: Innovation, Legal Risk, Critical Thinking, and Strategic Ops. When users rely on SmartInput, they bypass manual catalog navigation. The Brainstorm Reactor silently evaluates their natural language string, identifies the bilingual keyword footprint, and auto-embeds the most mathematically aligned methodology directly into the active prompt payload.

2. The Recipe System

A Recipe is a fully orchestrated cognitive pipeline enforcing structured thought over open-ended generation. Each Recipe orchestrates an exact sequence of strictly isolated steps involving:

• Methods: Formalized cognitive algorithms providing explicit step rules (e.g., TRIZ, Pre-Mortem).
• Workflow Vessels: Structural containers demanding exact formatting adherence.
• Execution Slots: The dedicated routing pipeline determining which upstream tier of computational model executes each unique step (SLOT_LOGIC, SLOT_CREATIVE, SLOT_SEARCH).

To enforce quality control, pipelines frequently execute steps internally before anything is presented to the user. Verification layers explicitly analyze previous generative nodes. If parameters fail, autonomous rerouting (Aletheia Loops) kicks the system back to the generation engine with new repair directives.

3. Workflow Vessels (WFS)

To produce highly useful artifacts, generative intent must be separated from structural formatting.

A fundamental flaw in traditional prompt engineering is commanding a model to "Think deeply about problem X, then format the result as a table." This merges the cognitive load with the formatting load. Within the Brainstorm Reactor, structural constraints are managed entirely by Workflow Vessels (WFS).

A Vessel formally dictates the ultimate payload container utilizing strict Markdown-driven rules. Regardless of whether a node is tasked with legal analysis, creative writing, or technical diagramming, the engine restricts the structural footprint of the output to match the defined Vessel parameters securely.

• Execution Separation: The methodology algorithm instructs the engine what to evaluate. The injected Vessel instructs the runtime how to construct the response.
• Deterministic Presentation: Invoking the Evaluation Vessel forces the model into creating strict dimensional grids with actionable justifications.
• System Translators: Highly specific Vessels, such as the Diagram matrix, actively force raw ideation protocols into pure executable schemas (Mermaid.js logic blocks or serialized JSON structures) capable of rendering native multi-dimensional UI overlays.