Agricultural extension is critical to post-conflict recovery in Liberia, yet sustained donor withdrawal has exposed significant dependency challenges to productivity. This study examined the level of post-donor agricultural extension service dependency, its effect on smallholder productivity, and the moderating role of institutional support capacity. A quantitative explanatory cross-sectional survey design was selected. This design enables systematic hypothesis testing and identification of associations within a representative sample at a specific point in time. Data were collected from 395 randomly selected smallholder farmers across Bong, Lofa, Nimba, Grand Bassa, and Margibi counties. Data were collected using a structured questionnaire measuring dependency, productivity, institutional support, and transition barriers. The instrument was validated by five agricultural extension experts, pretested among 40 respondents, and demonstrated acceptable reliability. Data were analysed using descriptive statistics, Pearson correlation, multiple regression, and moderation analysis. Findings showed a high overall dependency level (mean = 3.11), with behavioural dependency highest (mean = 3.38). Technology adoption continuity, defined as the sustained utilisation of donor-introduced practices beyond project closure, remained low (38.2%), indicating weak post-project sustainability. Dependency significantly and negatively affected productivity (β = −0.43, p < 0.001), explaining 37% of productivity variation and confirming H₁. Institutional support capacity significantly moderated this relationship (β = 0.19, p < 0.001), increasing explanatory power to 51%, thus supporting H₂. Major barriers included stipend expectations (mean = 3.52) and weak government extension coverage (mean = 3.34). The study concludes that strengthening institutional support systems is essential for sustaining agricultural productivity and extension system resilience in Liberia.