Thesis: OpenQodex makes the pre-push review stage a combined scanner-and-agent workflow.
Observed facts: The GitHub description says OpenQodex combines more than 100 scanners with a separate AI reviewer. It also says the reviewer uses an existing Claude Code or Codex login, without a separate key, and positions the review before push.
Mechanism: Scanner findings are investigated by a distinct AI reviewer using the developer's existing Claude Code or Codex login, with the review placed before push. This connects automated detection to investigation without claiming that the model replaces tests or human review. Why now: Coding agents are producing code inside developer workflows, making the timing and verification of review a practical design question.
First implication: Teams can evaluate this combined review workflow against their pre-push checks and credential policies. Evidence boundary: The supplied GitHub description establishes the project's stated design, not measured detection quality, comparative coverage, or cost savings. First watch condition: Look for published evaluation results showing which findings the AI reviewer correctly validates or rejects. New branch: No prior New Runtime record was supplied, so this release starts a distinct analytical record without assuming a related earlier analysis.