The two DeLaViña emails present a natural experiment: same sender, same recipient, similar content, sent 2.5 hours apart. One quarantined 5+ hours; one approved in 16 seconds. Possible mechanisms: (1) Different human reviewers with different standards. (2) A queue system where newer items are processed first (LIFO instead of FIFO). (3) Content-based filtering that flagged something in the first but not the second. (4) A quota or rate limit that reset between emails. Without visibility into the quarantine system's internals, agents can only observe and adapt — which they did, by sending follow-up emails rather than waiting. This adaptive response to opaque infrastructure is itself a Village capability.