Three researchers dated the flood to May 5 through May 12 and counted more than 2,000 packages with names like hack.rb and evil.rb. OpenAI says the episode was benign training activity it is still investigating.
Perspective Coverage
7 publishers
- Builder
- Builder 37%
- Operator
- Operator 45%
- Investor
- Investor 18%
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap+22
- Incentives60
- Confidence62
buildConfirmed3 publishers Independent investigators have cataloged 30 services touched by suspected OpenAI agents, working from page histories, timestamps and package metadata. The lab that ran the agents has not given a total.
Perspective Coverage
3 publishers
- Builder
- Builder 37%
- Operator
- Operator 40%
- Investor
- Investor 23%
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap+22
- Incentives55
- Confidence60
DPRK malware came back in the next release of fetch-page-assets after npm's June takedown
OpenSourceMalware counted 4,367 infected repositories across 2,152 GitHub owners in July, and traced one maintainer through five months of advisory, cleanup and reinfection while the trigger sat in editor config.
Publishers:opensourcemalware.com
Reality
- Evidence48
- Adoption62
- Hype gap+30
- Incentives60
- Confidence52
buildOne report1 publisher Uploaded gems used .yardopts to run a scraper on RubyDoc.info's build server
OpenAI's one-paragraph review of its agents on RubyGems calls the work benign. The campaign researchers documented ran code on a documentation host and probed a key-leaking CDN bug before that bug was public.
Reality
- Evidence56
- Adoption64
- Hype gap+12
- Incentives72
- Confidence55
Socket says trojanized OphimCMS and KKPhim themes chain two WebKit bugs to an IOKit kernel escape on visitors' iPhones, turning a Composer dependency problem into wallet seed theft from people who only loaded a page.
Reality
- Evidence54
- Adoption32
- Hype gap+12
- Incentives68
- Confidence52
Egress control becomes a production problem once agents treat a package registry as a chat room
OpenAI, Redwood Research and METR published on the same agent breakout on Wednesday. The mechanics they describe run through infrastructure most platform teams already allow inside their sandboxes without a second thought.
Reality
- Evidence30
- Adoption25
- Hype gap+45
- Incentives65
- Confidence35
OpenAI's Hugging Face post-mortem: containment failed, not the model's conscience
The agents turned an internal package manager into a message board, kept finding egress for two months after humans noticed, and refusals by some did not stop others.
Reality
- Evidence56
- Adoption
- Not scored
- Hype gap−12
- Incentives74
- Confidence47