Objectives We assessed associations between child stunting, recovery, and faltering with schooling and human capital skills in a native Amazonian society of horticulturalists‐foragers (Tsimane'). Methods We used cross‐sectional data (2008) from 1262 children aged 6 to 16 years in 53 villages to assess contemporaneous associations between three height categories: stunted (height‐for‐age Z score, HAZ<–2), moderately stunted (–2 ≤ HAZ≤–1), and nonstunted (HAZ>–1), and three categories of human capital: completed grades of schooling, test‐based academic skills (math, reading, writing), and local plant knowledge. We used annual longitudinal data (2002–2010) from all children (n = 853) in 13 villages to estimate the association between changes in height categories between the first and last years of measure and schooling and academic skills. Results Stunting was associated with 0.4 fewer completed grades of schooling (∼24% less) and with 13–15% lower probability of showing any writing or math skills. Moderate stunting was associated with ∼20% lower scores in local plant knowledge and 9% lower probability of showing writing skills, but was not associated with schooling or math and writing skills. Compared with nonstunted children, children who became stunted had 18–21% and 15–21% lower probabilities of showing math and writing skills, and stunted children had 0.4 fewer completed grades of schooling. Stunted children who recovered showed human capital outcomes that were indistinguishable from nonstunted children. Conclusions The results confirm adverse associations between child stunting and human capital skills. Predictors of growth recovery and faltering can affect human capital outcomes, even in a remote, economically self‐sufficient society.