We lost ₹22,845 crore to cyber fraud in 2024. A 206% rise from the year before. I want to take a moment to acknowledge something before I get into the idea — the people behind CERT-In, the cybersecurity researchers, and the platform safety teams are working hard. This isn't a criticism of their effort. This is a recognition that the problem has outgrown the current structure. Because here's what's actually happening on the ground: A malicious link gets flagged on WhatsApp. It spreads freely on Instagram. Gets reshared on X. Someone's grandmother in a tier-3 city clicks it at 11 PM. Her life savings — gone. No warning. No safety net. Nothing. This isn't a hypothetical. This is Tuesday in India. The root issue isn't effort. It's fragmentation. Every platform runs its own detection system in isolation. Meta has its own. Google has its own. X has its own. They don't share intelligence. A link that's been confirmed malicious on one platform can take hours — sometimes days — to get flagged on another. And with AI now generating phishing links that are indistinguishable from legitimate ones, at unprecedented speed and scale, those hours cost lives and livelihoods. The solution I'd like to put forward is straightforward in principle: Build a single, open-source malicious link detection API. Jointly maintained by CERT-In, Meta, Google, X, and the broader developer community. One shared threat intelligence layer. Universal. Real-time. Sub-second response. Zero licensing barriers. Every platform, every app, every developer in India plugs into the same engine. A link confirmed malicious anywhere gets flagged everywhere — simultaneously. CERT-In already coordinates with 1,400+ organizations for cyber drills. The institutional framework exists. What's missing is a shared technical standard that sits underneath all of it. I'm grateful for every person working in this space. And precisely because of that gratitude — I think they deserve better infrastructure to work with. This is a public good. It should be built like one. Would love to hear from developers, policy folks, or anyone in platform safety who's thought about this. Is anyone already working on something like this? What are the real blockers?