Viral tweet claims LLM found a Redis 0-day — and dangles a malware-bait link

A Hacker News-surfaced tweet claims that an LLM called “Kimi K3” independently discovered a zero-day in the latest Redis server and used it to write a working exploit, allegedly taking 27 minutes across 32 coordinated agents. The poster frames it as the first model both capable of and willing to author real exploit code, and links to a GitHub repo said to contain the proof. None of this is verified: there’s no confirmed Redis advisory tied to the claim, no independent reproduction, and “Kimi K3” is not an established, documented model. Treat the whole thing as an unproven social-media assertion until a named CVE, vendor confirmation, or reproducible write-up appears.