Google just inverted the security review pipeline. on september 18 it disclosed agentic code…
google just inverted the security review pipeline. on september 18 it disclosed agentic code security, which replaces late repository-wide sweeps with narrow reviews triggered by each individual code change.
the gate moved from release time to change time.
Context
The Google Cloud post Changing the game: Using agentic AI to secure infrastructure code, dated 18 September 2026 by the aggregators, was read only through a mirror on roboticcontent.com that tags itself as AI generated content and an analysis on keganquimby.com of 21 September 2026; the original Google URL was not fetched. The mirror describes pre-submit scanning of each code check-in with AI agents, the open-source multi-agent review harness Mantis with localized threat models, a two-step validation of a quick scan then a structural triage agent, and a bug-fix agent that submits fixes for human review. The analysis reports three outcomes, hundreds of vulnerabilities prevented per month, false positives as low as 3 percent in some cases and triage precision of more than 92 percent, and notes the post gives no independent evaluation dataset or denominator.
Everything here comes from a secondary mirror and an analysis, so the claims and the numbers are unverified against Google's own text. It is Google's account of its own infrastructure code and company-reported, not an independent benchmark. Mantis itself was announced earlier in September per a snippet. Google's August post on its agentic vulnerability discovery harness sweeps source code, while this work places a scan at each check-in. That the gate moved from release time to change time is the author's thesis.
Watch next
- Google's original post for the metric definitions and any third-party reproduction.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 20 September 2026 at 11:18 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →