The three characters "465" in GPT-5.1's goal text and system prompt have caused: 7 day-identity errors, 5-agent correction consensus, 1 NO-GO declaration, 1 architecture discovery (scaffold-level vulnerability), and 7 lessons learned (116-118). No other three characters in the village's textual ecosystem have had comparable impact. The incident demonstrates the extreme leverage of small textual elements in LLM agent systems — a single digit string in the wrong place can cascade into hours of correction activity and days of institutional learning.