Add first-contact, impact, and commercialization guides

[alanthssss]

Summary

• add equal English and Chinese zero-background introductions
• place a plain-language entry point before milestone details in both READMEs
• add a four-step project story and audience-specific routes to both website languages
• clarify that M14 is the promoted policy, M15 is the decision workflow, and M16 is the engineering milestone
• explain current practical value, transfer scenarios, customer hypotheses, strengths, weaknesses, and commercialization paths
• define evidence gates from customer interviews through transfer, pilot, and productization

Why

The existing presentation assumed readers already understood the project domain and did not clearly distinguish current value from future commercial potential. New visitors—including technically experienced readers who are new to multi-agent learning—were asked to interpret milestones and specialist terms before learning the underlying cooperation problem.

This change makes the reading order progressive: problem first, practical meaning second, evidence and specialist detail afterward. Commercial claims are intentionally bounded: SweetGold is an open-source research and engineering project today, not a validated warehouse, robotics, or governance product.

External context

The impact analysis is grounded in primary sources from PettingZoo/Farama, NVIDIA Isaac Lab, Google DeepMind multi-agent research, and the NIST AI Risk Management Framework. These sources establish the relevance of multi-agent simulation and repeatable AI evaluation; they are not presented as validation of SweetGold in those industries.

Validation

• 53 Python tests passed, including bilingual document parity
• 3 server-rendered website tests passed
• production website build passed
• GitHub Pages export passed
• diff whitespace check passed