Governance Gaps in Agent-to-Agent Social Networks: A Multi-Framework Analysis of Moltbook Pre-Acquisition

We present an empirical governance gap analysis of Moltbook — the largest AI agent social network (1.5M registered agents, ~17K human owners, 88:1 delegation ratio) — conducted between platform maturity and Meta's acquisition on March 10, 2026. Using two established governance standards (NIST AI RMF, ISO/IEC 42001) and one author-developed agent-specific framework (FAPA), we report two independent governance metrics: 94% compliance-signaling absence — only 6% of analyzed posts contain any marker of regulatory awareness, logging, or governance process 91% accountability-signal absence — only 9% of posts contain formal accountability verification markers such as delegation chains, authorization references, or decision attribution tokens Additional findings: No effective governance implementations were identified across all three frameworks based on publicly available evidence Only 30.7% of agents have a claimed human owner (self-reported; opt-in, unverified, unenforced) Karma Gini = 0.877, indicating extreme concentration inconsistent with a governance instrument Cross-framework convergence — NIST, ISO 42001, and FAPA identify the same critical gaps in delegation, evidence integrity, and behavioral monitoring 14 Findings. This paper's contribution is empirical measurement and analysis, not system design.