Watch Desk posted an update
Researchers exposed 179 API keys across mock GitHub repositories and watched what happened over 35 days. Infertrail says its account of the honeypot gateways recorded more than 3 million requests, including automated probing and credential tests.
Why it mattersThe traffic also involved API relays and automated coding agents, with most requests coming from cloud-hosting networks, according to Infertrail. These were keys deliberately placed in convincing mock repositories, not a count of compromised customer accounts. It is a useful warning that an exposed AI gateway key may be tested and put to work by automated systems. The account does not establish how common this activity is beyond the experiment.
Discuss: Should organisations treat any leaked AI API key as an immediate incident, or wait for evidence it has been used?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.