Causal Localization of Agent WANDERING to Edge-Layer L11: The Right Locus Is Still Not a Rescue Lever

Three causal tests of the Tool-Entropy WANDERING mechanism hypothesis (mid-layer verdict consolidated, edge-layer alignment fails) on Qwen3.6-27B SWE-bench Pro trajectories. All three are null on rescuing WANDERING: a forced-finish counterfactual (Fisher p=0.71), an always-on L55 SUCCESS-donor hook (Fisher p=1.00), and a norm-matched L11 alpha-sweep at the strongest detector locus (paired McNemar p=0.73). The one robust effect is the opposite of a rescue: high-magnitude L11 injection destabilizes the model into invalid tool calls (12/20). A load-bearing methodological finding: WANDERING is not run-stable at temperature 1.0 (the same instances flip finish 7/20 with no intervention), so every intervention test must be paired and the unpaired 0/20 baseline manufactures a false positive. Companion to Tool-Entropy Collapse (DOI 10.5281/zenodo.20368601).