Seven-Stage Mind Engine

The Specific Problem It Addresses

Many AI agents jump from an ambiguous user question directly to a single answer, skipping key assumptions, missing constraints, and under-checking risks. mind-engine turns general thinking into an executable framework: the agent diagnoses the problem, matches analytical models, asks clarifying questions, and only then compares multiple reasoned options.

How It Works and Where It Fits

The skill is organized around a seven-stage reasoning process:

• Problem diagnosis: classify the question as factual or normative, deterministic or probabilistic, one-shot or recurring, and identify stakeholders, hidden assumptions, and cognitive biases.
• Model matching: select primary and auxiliary models such as game theory, Bayesian updating, systems thinking, first principles, and decision science.
• Dialogue exploration: ask questions about goals, constraints, known information, involved parties, time horizon, risk tolerance, and prior experience.
• Hypothesis generation: produce at least three distinct paths, each tagged with conditions, possible outcomes, key risks, and methodology source.
• Exhaustive verification: check each hypothesis for ergodicity, stepwise validity, skin in the game, recursive traps, worst-case exposure, and antifragility.
• Recommendation and consolidation: deliver recommendations with reasoning chains, then record user preferences, constraints, and observed method limits.

It is useful for decision analysis, clarity-seeking, and multi-option comparison. It is less suitable for fast factual lookup, rigid procedural tasks, or situations that require one immediate answer. Without a custom knowledge base, the skill uses a generic model table; with one, users can replace or append their own methodology index.