SecurityThe briefing
AI-generated illustration
AI models now discover and exploit vulnerabilities in open source libraries within a day of disclosure. Enterprises must shift from 90-day patching cycles to hours while paying for maintained packages.
The median time from disclosure to first exploitation fell from roughly a year in 2021 to one day in 2026, with almost 80 percent of exploits now landing on or before public disclosure, according to J.P. Morgan Asset Management analysis cited in the report.
Anthropic and Project Glasswing partners reported more than 10,000 new high- and critical-severity zero-days in the first month of their effort, while only 75 of 530 high- and critical findings sent to open source maintainers by May 2026 had been patched.
Datadog February 2026 data showed the median dependency 278 days behind its latest major version, leaving old code exposed to one-day exploits that models such as Anthropic Mythos and OpenAI GPT-5.5 can now locate at scale.
Professional open source backed by vendors or staffed foundations is presented as the required shift, because single-maintainer packages and abandoned code can no longer be relied upon when the exploitation clock runs in hours rather than months.
Builders evaluating dependency pipelines should test routing builds through an internal proxy such as JFrog Artifactory or Sonatype Nexus that points at a supported source like TrueSource. In an illustrative scenario, a team running 50 critical dependencies would first map which packages have identifiable maintainers and which are single-person efforts, then configure the proxy to block known-vulnerable versions and serve the nearest patched release automatically. The evidence does not establish whether this proxy change alone would meet regulatory audit requirements or how quickly maintainers of the chosen supported sources respond to new zero-days.
Operators responsible for technology risk should measure time from disclosure to deployed fix for their top 50 dependencies and maintain a list of maintainers for each critical package. In an illustrative scenario, a financial services platform would run the proxy change described above without altering developer workflows, record the resulting reduction in days from disclosure to production, and compare that metric against the one-day exploitation window reported in the J.P. Morgan note. The supplied evidence does not establish the cost of switching to paid professional open source packages or the exact eligibility terms for any particular supported source.
Identify the top 50 dependencies in production, record the organization or individual listed as maintainer for each, and note the current version age relative to the latest release. Re-run the same inventory after configuring an internal artifact proxy to a supported source and measure the change in days from disclosure to deployed fix; the difference informs whether the proxy approach narrows exposure enough to justify continued use of the current open source mix.
Source material behind this briefing. Our practical implications are Ashlr’s interpretation.
This briefing was prepared with AI assistance from the linked sources. The analysis may include interpretation; consult the original sources for complete context. Cover artwork is illustrative.
Anthropic's Claude Code v2.1.311 introduces CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS, removing the final guardrails on commands like rm. The change follows the earlier --dangerously-skip-permissions flag and bypassPermissions mode.
SecurityThe briefing
AI-generated illustration
AWS made its Continuum agent generally available. The service finds, validates, and proposes fixes for code vulnerabilities in repositories while checking the AWS environment.
SecurityThe briefing
AI-generated illustration
Nothing merged in Ashlr's public repositories on October 21. Here is the idea behind a phrase that runs through recent pull requests: value-free reports.
