ContextFlow: Hierarchical Task-State Alignment for Long-Horizon Embodied Agents

TL;DR

ContextFlow is a framework that maintains task-level consistency in long-horizon embodied agents by explicitly aligning planning, monitoring, and execution stages, improving robustness and transparency.

Contribution

It introduces an inspectable alignment framework that explicitly represents task stages and applies scoped updates to mitigate task-state failures.

Findings

Experiments show improved task success rates with ContextFlow.

The framework enables explicit diagnosis of task-state failures.

Demonstration traces illustrate effective mitigation of recurring issues.

Abstract

Long-horizon embodied agents increasingly delegate navigation, search, approach, and manipulation to specialist executors. As these executors become stronger, the main bottleneck shifts from local skill execution to maintaining a coherent task frontier across planning, monitoring, memory, and execution. We study task-state misalignment, a task-level consistency failure in which the planner's active stage, runtime evidence, remembered context, and delegated executor no longer justify the same next-step decision. This failure can lead to unsupported handoffs, stage lock, executor-context mismatch, and unnecessary replanning. We propose ContextFlow, an inspectable alignment framework that represents stages as explicit contracts, converts runtime observations into evidence packets, and applies scoped updates including continue, refine, transfer, promote, and repair. ContextFlow keeps specialist executors responsible for local closed-loop control while making task-frontier alignment explicit and auditable. Experiments and demonstration traces on long-horizon embodied tasks illustrate how evidence-grounded scoped updates diagnose and mitigate recurring task-state failures.