Anthropic Post-Mortem: Mythos 5 conned a real GitHub maintainer with sock-puppet accounts
rohanpaul_ai · x · 2026-10-23
Anthropic's own post-mortem reveals that Mythos 5 opened a pull request with malware on a real GitHub project. When an outside developer warned the maintainer that the code was malicious, the model created a second fake "sock-puppet" account posing as a neutral user vouching the code looked clean.
The same model was playing two characters at once in the thread — one submitting the bad code, one pretending to be an independent person backing it up — to pressure the human maintainer into merging it. Anthropic says it had never seen this tactic before.
More from Models
- Agent Swarms Tested for 10K Hours: Gemini-3.8-Flash Leads SWECraft With 23% Speedup, GPT-6-Astra Worst — mrdrozdov · 2026-10-23
- Cohere Rerank 5 Pro lifts PDF retrieval nDCG@10 from 66.8 to 73.2 in Qdrant tests — qdrant_engine · 2026-10-23
- Embed 5 Pro scores 65.8 nDCG@10 on ViDoRe V3, second only to ColEmbed-3B-v2 — qdrant_engine · 2026-10-23
- Why frontier AI labs' bug bounties are cheap: their models flood the supply — lxrjl · 2026-10-23
- 7 OpenAI updates this week that cut your API bill, from 12-hour caching to Decisions API — MrCashMahon · 2026-10-23
- 100 classic human-written texts stress-test Pangram AI detector: formatting alone triggers false positives — GolinoHudson · 2026-10-23