Instruction Bleed: Cross-Module Interference in Prompt-Composed Agentic Systems
For practitioners building prompt-composed agentic systems, the paper establishes cross-module interference as a measurable and overlooked failure mode that current evaluation practices miss.
The paper identifies and formalizes compositional behavioral leakage (CBL), a failure mode in prompt-composed agentic systems where editing one module silently affects others due to shared context windows. Experiments on a deployed job-evaluation agent (144 trials) show a detectable but sub-threshold effect (Cohen's d = 0.63) that does not flip recommendations but could compound over many decisions.
Practitioners of prompt-composed agentic systems report a recurring failure mode: editing one prompt module silently shifts the behavior of others despite no shared variable or executable dependency. We formalize this as compositional behavioral leakage (CBL): interference between modules sharing a context window. CBL is enabled by architectural non-isolation: transformer self-attention provides no formal boundary between concatenated modules. We probe CBL on a deployed job-evaluation agent (Claude Sonnet 4.6, 144 trials) through a reusable three-channel protocol that perturbs non-focal modules along volume, content, and form. Only the content channel produces a detectable paired effect (Cohen's d = 0.63, bootstrap 95% CI excluding zero); no recommendation flipped -- a sub-threshold regime invisible to standard QA but compounding across the thousands of decisions a deployed agent makes. CBL is orthogonal to known agent-failure axes (adversarial injection, cognitive degradation, multi-agent fault propagation, privacy leakage). We contribute an operational definition, a reusable protocol, a falsifiable prediction set, and a system-class characterization, establishing cross-module interference measurement as a requirement for prompt-composed agent evaluation.