RATIONALE: Psilocybin is a potentially paradigm-shifting depression intervention. We conducted a systematic review and meta-analysis of psilocybin-for-depression randomized controlled trials (RCTs). OBJECTIVES: Systematically assess harm reporting, risk of bias, action mechanism specification, and incremental therapeutic effect sizes in the psilocybin-for-depression RCT literature. METHODS: Assessed databases included PsycINFO, CINAHL, Embase, Medline, Web of Science, and Scopus. Search terms "Psilocybin" or "Psychedelic" were paired with "Depression", and "Randomized Controlled Trial" or "RCT". RESULTS: We identified k = 9 RCTs (k = 10 subgroups) involving n = 602 participants (56% psilocybin). Five studies had low/very low harm quality reporting, opposed to two with high. Most studies demonstrated a high risk of bias. Therapeutic mechanisms of action (MoAs) were discussed in varying detail but rarely assessed in original publications. Psilocybin was moderately superior to controls at reducing depression (g = 0.62; 95% CI = 0.27, 0.98). Effects were heterogenous (τ = .47). Smaller studies evidenced stronger effects that favored psilocybin (Egger's b0 = 3.63, p = .014). Almost all studies documented financial conflicts of interests. CONCLUSION: Psilocybin demonstrates significant depression reduction relative to controls. However, researchers, clinicians, and stakeholders should consider several contextual factors. Effects were moderate and attenuated in larger and better-controlled studies. Harms reporting and risk of bias was high, though partly driven by unique challenges of psilocybin research. MoAs were variably specified but rarely assessed; suggesting it is unclear how depression is reduced. We advise researchers conduct RCTs with active control conditions, larger samples, and include MoA assessments. Independent RCTs from researchers without financial conflicts of interest are needed.