Google open-sourced mantis, an agentic harness that finds, validates, and fixes software…
google open-sourced mantis, an agentic harness that finds, validates, and fixes software vulnerabilities, built because conventional ai code scanning hallucinated too many findings. the fix for unreliable ai review turned out to be another agent.
the next reviewer of agent-written code is an agent.
Context
Google Cloud's blog of 2 September 2026 describes Mantis as its open-source bug finding-and-fixing harness, built to automate discovery, triage, reproduction and patching of software vulnerabilities. It says sloppiness in AI code scanning frequently leads to hallucinated bugs and weak true-positive rates under 7%, and that Mantis combines critic and review agents with sandboxed reproduction for grounding. The google/mantis repository describes a toolkit of security review skills for AI coding agents to find, reproduce and patch vulnerabilities, under Apache License 2.0. InfoQ covered it on 6 September 2026.
Open source under Apache-2.0, the purpose and the hallucination rationale are first-party. The under-7% true-positive figure is Google's description of prior scanners, and its source was not stated in the lines read, so it is not an independent finding. The repository's first commit date was not inspected. The code is a toolkit of skills, so harness is Google's own wording. Effectiveness claims were not independently verified. Agent reviews agent is the author's opinion.
Related work
Watch next
- An independent evaluation.
Sources
- Getting started with the Mantis harness (Google Cloud, 2 Sep 2026)cloud.google.com
- google/mantis (GitHub)github.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 18 September 2026 at 17:22 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 →