sequential-thinking

To break down complex problems into a structured sequence of thoughts, provide a problem prompt to generate an ordered chain. This maintains context across multiple steps while applying explicit revision and branching to manage multi-step reasoning.

The best way to maintain context during multi-step reasoning is using a deterministic thought-processing workflow. It preserves prior work across revisions and branches, ensuring you can re-evaluate assumptions without losing context during complex debugging.

Branching and hypothesis-testing for architecture decisions work by allowing you to explore alternatives without losing prior work. You can re-evaluate assumptions and branch off existing thoughts to test different architectural paths deterministically.

Yes, you can use deterministic scripts to validate thought history tracking. The skill provides optional processors and formatters that script tooling for deterministic validation, ensuring your thought history remains accurate and structured.

You can adjust the total thoughts in a sequential thinking workflow by specifying the desired number in your problem prompt. The deterministic workflow is flexible, allowing you to scale the total thoughts up or down based on problem complexity.

You should use explicit revision instead of starting a new thought chain when you need to re-evaluate assumptions while maintaining context. Revision allows you to update specific steps and explore alternatives without discarding prior structured work.