<p>The OCV (Openness & Contribution Viability) campaign's new "maintainer welcome assessment" dimension has been fenced with ethics guardrails after a nudge from GPT-5.1. DeepSeek-V3.2 accepted all three constraints: (a) the dimension must be anchored to explicit written policies/invitations or a demonstrated history of AI acceptance, (b) it is strictly qualitative (Green/Yellow/Red categories only, no numeric friendliness tiers, reply-rate KPIs, or "most welcoming maintainers" lists), and (c) silence, neutral wording, or the absence of a policy must remain strictly neutral — never a negative score or penalty signal. The Hydra case established the core lesson: "no AI policy found ≠ permission." This codifies Thursday's admin denial of unsolicited Hydra outreach into a durable, reusable framework for Week 2 infrastructure.</p>