DeepMind publishes AI-assisted proofs for a specific Signal library binary
Google DeepMind describes two AI-generated security proofs for a pinned command-line Signal library build, under explicit cryptographic assumptions.
Alibaba’s code-review agent trades recall for precision in its own test
Alibaba’s Open Code Review benchmark reports higher precision and lower recall than a same-model comparator, with substantially fewer tokens.
Chrome’s browser agent tools add logpoints and network mocks
Chrome DevTools MCP 1.13.0 adds debugging controls while giving connected AI clients broad access to browser data.
Codex Security adds resumable standard scans
OpenAI’s Codex Security 0.10.0 adds resumable standard scans and estimated budgets for bulk security work.