OmniContact: Chaining Meta-Skills via Contact Flow for Generalizable Humanoid Loco-Manipulation

Learning long-horizon humanoid loco-manipulation poses a dual challenge: it requires not only the robust execution of meta-skills but also their seamless, closed-loop chaining equipped with autonomous recovery. Existing approaches remain limited: explicit humanoid-object interaction representations offer precision but are notoriously difficult for high-level planning, whereas implicit skill embeddings are compact but lack the interpretability required for reliable composition. We propose \ours, a hierarchical framework centered on $\mathbf{\text{contact}}\mathbf{\text{flow}}\mathbf{\text{(CF)}}$, a compact representation consisting of key body trajectories and time-series binary contact signals. Leveraging this shared interface, our low-level policy $\mathbf{\text{CF-Track}}$ learns a unified library of loco-manipulation skills, while our high-level module $\mathbf{\text{CF-Gen}}$ heuristically synthesizes future contact-flow sequences. To support this setting, we additionally collect the OmniContact dataset, a MoCap-based HOI corpus for humanoid loco-manipulation (Appendix~\ref{sec:dataset}). Together, they enable robust execution, autonomous failure recovery, and flexible composition of meta-skills for long-horizon tasks. Experiments show that OmniContact achieves $98.7\mathrm{\%}$ success on $\mathit{\text{Carry}}\mathit{\text{Box}}$ and $76.5\mathrm{\%}$ on $\mathit{\text{Push-Stack}}\mathit{\text{Boxes}}$, outperforming prior baselines by average margins of $40.9\mathrm{\%}$ in meta-skill and $66.5\mathrm{\%}$ in skill chaining. Besides, our framework naturally integrates with VLMs for semantic task decomposition, enabling complex, semantically grounded loco-manipulation behaviors, such as arranging scattered boxes into a heart shape.