Working through a problem

Start with a hypothesis or a hard problem

A load-bearing hypothesis supports an important decision. Write it in one sentence, then name the decision that depends on it. If you cannot yet state a hypothesis, describe what is stuck and who experiences the problem.

Form problem statements and smaller questions

Separate observations from explanations. Break the problem into questions that can be investigated: where the delay occurs, what each team needs and which constraints are assumed. Keep each part connected to the original hypothesis.

The shape can change. A subsection may become a distinct problem; two questions may turn out to share one dependency. Do not force the work into the first outline.

Use Theo to guide the inquiry

Theo contributes methods, frameworks and questions to the investigation. Ask what kind of progress you need before choosing a method: clarify the need, challenge an assumption, compare explanations or design a discriminating test.

Theo reads problem shapes, not content. The plugin uses this method guidance while working with your project context. Read the Theo connection and data boundary.

Examine relationships and uncertainty

Ask which evidence supports the hypothesis, what contradicts it and which downstream decisions would change if it failed. An unknown should remain visible. Do not turn an uncertainty into a probability just to make the map look complete.

A new relationship may reveal an opportunity: for example, changing the information supplied at the handoff rather than adding capacity. Treat that as a candidate explanation until it is tested.

Choose the next test yourself

The system can propose a next step. You decide whether its evidence and consequences justify taking it. At a material approval gate, review the proposal and choose whether to approve, reject or defer. Keep your reason with the decision.

A eureka moment changes what you can see. It does not validate the opportunity or authorize an action.

Keep gaps and failed tests visible

When a source or method is unavailable, keep the gap explicit. If Theo cannot be reached, follow the connection checks. Do not treat an unsourced answer as a sourced method.

Your first working session

1. Bring the hypothesis and one real piece of context.
2. State the problem it creates and the decision it supports.
3. Identify one assumption or relationship worth investigating.
4. Choose an evidence-producing next step and its success criterion.
5. Return the result to the same DataRoom and review what changed.

Between sessions, use /mos:status to review the room. Progress may mean a better question or a rejected explanation; it need not mean the whole problem is resolved.