Open-source maintainers report AI agents turning bug rumors into exploits within minutes, straining disclosure and CVE processes
2026-08-28
Anil Madhavapeddy (Cambridge computer science professor and OCaml core maintainer), writing on his blog as relayed by Simon Willison, reported that security patches shared for discussion on OCaml projects are now drawing automated exploit probes within about ten minutes, down from the previous norm of days before an issue or release. He attributed this to modern AI coding agents being effective enough that a mere rumor of a bug gives them enough information to find and exploit it, and said he demonstrated this himself with his own agents, switching to DeepSeek V4 Pro after Claude Fable refused the task. Rclone maintainer Nick Craig-Wood confirmed the pattern in Hacker News comments: rclone received about 20 security disclosures via GitHub in its first 10 years, but more than 40 in the last month alone, with roughly 75% containing 'a nugget of something which needs looking at'; he said GitHub's CVE assignment turnaround has slowed from 2-3 days to 3-4 weeks under the volume, forcing rclone to ship point releases marked CVE-PENDING.
Significance 3: Concrete, named, cross-project numbers (rclone's 2x-plus disclosure volume, roughly 10x-slower CVE turnaround) show AI-driven vulnerability discovery straining open-source security infrastructure industry-wide, but it rests on one curator's relay of named practitioners' public statements rather than a primary report, so held at 3 pending broader corroboration.
Implications · machine-drafted, not owner judgment
If bug rumors reliably convert to exploits within minutes, existing responsible-disclosure embargo norms (days to weeks) are already obsolete for widely-deployed open-source projects, forcing maintainers to either patch and ship faster with less review or adopt AI-assisted triage themselves — Craig-Wood says rclone is already doing the latter. The reported model-level split (Claude Fable declining the exploit-development request; DeepSeek V4 Pro complying) is a single anecdote but a concrete, checkable data point on differing refusal postures for offensive-security tasks across labs, worth tracking as more cases surface.
- Whether other major open-source projects beyond rclone and OCaml report similar disclosure-volume spikes
- Any formal proposal to reform CVE or embargo timelines in response
- Further documented cases of models differing on refusal for exploit-development requests
Sources
- curator Just a rumour of a bug is enough to find a security exploit these days retrieved 2026-08-29