OpenAI · 2026-11-12 · notable
Codex Security 0.10.0 — OpenAI's scanner can resume scans and cap costs
In short
Codex Security 0.10.0 lets interrupted scans resume from a checkpoint, adds a shared cost budget for bulk scans, imports ArmorCode SAST findings, and explains why a blocked patch failed. The repo is trending on GitHub today.
OpenAI · 2026-11-12 · 1 source checked
OpenAI's open-source vulnerability scanner gets resumable scans, cost budgets and clearer patch failures.
Key specs
| GitHub stars | 12,244 |
|---|
What is it?
Codex Security 0.10.0 is a new release of OpenAI's Apache-2.0 CLI and TypeScript SDK for finding, validating and fixing security vulnerabilities. Interrupted Standard scans can now continue from a saved checkpoint with `codex-security scans resume SCAN_ID`, and a `--max-total-cost` flag sets one estimated budget across bulk or component scans.
How does it work?
Standard scans now start with the vulnerability audit and record the file count from the scan's inventory. When an automatic patch is blocked, the result says whether files changed, what stopped the fix and what to do next. Saved ArmorCode SAST findings can be imported into the local workbench, and SDK callers can publish selected findings to Cloud.
Why does it matter?
Long scans of large repos no longer have to restart from zero after a crash or timeout, and teams scanning many repos can set a spending limit up front. The release notes warn that in-flight work can go over the estimate, and MCP registration scripts need the new `--all` or `--agent NAME` flags.
Who is it for?
security engineers and teams running AI code scans in CI
Try it
npm install --global @openai/codex-security && cs scan .