The Irreversibility Threshold: When Error in LLM Agent Pipelines Outruns Human Correction

Large language model (LLM) agents are increasingly composed into pipelines, in which each agent conditions on the output of the one before it. We show that this decomposition pattern carries a structural failure mode: as a multi-agent pipeline deepens, semantic error does not merely accumulate, it compounds. We develop a formal model of error propagation in linear LLM agent pipelines and give the accumulated error a closed form, the cascade equation, from which we derive a geometric lower bound with cascade ratio $r>1$. Two thresholds follow in closed form: a critical depth $m^{\ast }$, the deepest pipeline that stays within an error tolerance $\theta$ without intervention, and an irreversibility threshold $m^{†}$, beyond which no bounded human validator, modelled as a gate $h=\alpha \beta$ that combines domain expertise and provenance traceability with achievable capacity $\bar{h}<1$, can restore the output to within $\theta$. These partition pipeline depth into safe, recoverable, and irreversible regimes. The model reframes reliability, rather than per-agent capability, as the binding constraint on deep agentic workflows, offers a candidate, testable structural account of why many multi-agent generative-AI deployments fail to deliver value, and proposes $m^{†}$ as a candidate pre-deployment audit trigger for human-oversight requirements such as the EU AI Act. Empirically, two exploratory pilots and two pre-registered, replicated mechanism experiments map the regimes. Relaying facts under a behavioural token budget degrades output by erasure with negligible alteration in all three models tested. Processing steps inject distortion net of a gold-state floor, whereas the same models' relay distortion is indistinguishable from zero in one model and one to two orders of magnitude smaller in the other. A pre-registered coupling assumption, that compression injects distortion, is not supported in relay; in processing it is supported in one of two models, whose fitted parameters place the worst-case handoff at an interior compression, and not in the other. We report the failed calibration gates and an instrument-precision sensitivity that lowers the processing estimates without changing their ordering, except for one model at the value-only extreme of the measured precision interval. Following emerging usage, we call this failure mode a hallucination cascade, and give it, to our knowledge, its first closed-form account.