What problem does it solve? Designing a correctly-specified Markov null model for topological data analysis is error-prone: researchers omit the Markov order, under-power permutation counts, or choose statistics that are invariant under the null, invalidating p-values. This Skill enforces the P01-B mandate that Markov order k is always explicit and generates a complete, correctly-parameterized permutation test scaffold. ## Core Features & Use Cases - Null Model Scaffolding: Generates the full permutation_test_trajectories call with explicit markov_order, statistic, permutation count, landmarks, and seed parameters. - Memory Ladder Guidance: Maps the five-rung null hierarchy (label_shuffle, order_shuffle, markov k=1, markov k=2, stratified_markov1) so batteries include proper negative controls. - Design Error Detection: Flags common mistakes such as missing markov_order, under-powered permutation counts, landmark reuse, and statistics centered on the null's sufficient statistic. - Use Case: A researcher testing whether trajectory topology exceeds second-order Markov prediction runs /markov-null-design markov 2 to get the complete function call, results file naming convention, and Computational-Log logging stub. ## Quick Start Ask the assistant to scaffold a Markov-2 null model permutation test for your trajectory TDA pipeline with explicit seed and permutation count.
