1. Theory suggests that non-trophic interactions can be a major mechanism behind community stability and persistence, but community-level empirical data are scarce, particularly for effects on species interactions mediated through changes in the physical environment. 2. Here we explored how ecosystem engineering effects can feed back to the engineer, not only modulating the engineer’s population density (node modulation) but also affecting it’s interactions with other species (link modulation). 3. Gall induction can be viewed as ecosystem engineering since galls serve as habitat for other species. In a community-level field experiment, we generated treatments with reduced or elevated ecosystem engineering by removing or adding post-emergence galls to different plots of their host plant in the Brazilian Cerrado. We tested the effect of post-emergence galls on the galler, as well as on the galler-parasitoid and galler-aphid interactions. 4. The manipulation of post-emergence galls had little effect on the galler - abundance and survivorship were not affected, and gall volume changed only slightly - but modified interactions involving the galler, parasitoid wasps, and inquiline aphids. Aphid inquilines negatively affected density-dependent parasitism rates (interaction modification) likely by killing parasitised galling larvae. Post-emergence galls interfered with aphid inquilinism - likely by the provision of alternative habitat for aphids – and thus interfered with the negative effect of aphids on parasitism (modification of an interaction modification). 5. This work is one of the few studies to demonstrate experimentally the role played by environmentally-mediated interaction modification at a community level in the field. Moreover, by manipulating a species’ ecosystem engineering effect (post-emergence galls) instead of the species itself, we demonstrate the novel result that populations can be regulated by non-trophic effects initiated by their own activities that alter their interaction with other species. This reveals that indirect interactions mediated via the environment offer new pathways of feedback loops for population regulation. Our results indicate that interaction modification has the potential to be a key regulatory mechanism underlying interaction variation in nature, and play a major role in community structure, dynamics, and stability.