Skip to content
worth noting Productivity

GitHub Copilot code review: adaptive review depth, secret detection, and context from other repositories

clearly official source

GitHub Copilot code review now offers Adaptive (automatic selection of review depth based on PR risk), detection of exposed secrets, and context from public repositories. All of this is in public preview.

GitHub has expanded the Copilot code review tool with three new features in public preview. Adaptive automatically selects the depth of the review level (Lite, Balanced, or Max) based on the complexity and riskiness of the pull request, so routine changes get a quick check while risky changes (for example in authentication logic) get a more detailed analysis. Users can enable Adaptive for an individual pull request or as a default setting at the personal, repository, organization, or enterprise level; the default level remains Balanced, and Adaptive will never exceed the maximum allowed level set for the enterprise, organization, or repository.

The second new feature is secret detection in pull requests. According to GitHub, it combines deterministic methods (regex, over 500 types of secrets developed in collaboration with more than 150 partners as part of the secret scanning program) with detection using ML/LLM models for secrets without a recognizable format. It is enabled in the repository or organization settings, under Copilot, Code review.

The third feature is cross-repository context, which allows the review to read related public repositories and documentation to detect code in other repositories that depends on the given change; this feature is in public preview and, for now, only for public repositories. All three features are available for the Copilot Pro, Pro+, Max, Business, and Enterprise plans. In addition, the company has made Repo Insights available for tracking how review feedback is used, and Copilot code review for Azure Repos has moved from public preview to general availability.

What changed

Why it matters

Automatic selection of review depth based on risk saves reviewers time on routine changes while ensuring more thorough checks for sensitive code, without a person having to set it for each PR individually. Secret detection directly within code review reduces the risk of credentials or API keys ending up in the repository's main branch before anyone manages to remove them.

Two audiences, two different impacts

What this means

01

For individuals

Developers using Copilot code review can leave the choice of review level (Lite/Balanced/Max) to Adaptive instead of manually setting it for each pull request, and rely on automatic alerts about exposed secrets before merging.

What to do Enable or try Adaptive as the default review level for your own pull requests.
More practical updates →
02

For a business

Companies with Copilot Business or Enterprise plans can use secret detection and cross-repository context to reduce the risk of credential leaks from pull requests and improve the consistency of code review across public repositories; they also now have access to Repo Insights for tracking how review feedback is used.

Risks and compliance
What to decide Check in the repository or organization settings whether secret detection is enabled in Copilot code review.
More business impacts →
Adaptive effort code review GitHub Copilot LLM secret detection

Check the original

Event sources

clearly official source · 1 publisher, 0 independent. We count feeds from the same owner only once.

1
GitHub Copilot Changelog primary source · first detected Copilot code review: Adaptive effort, secret detection, and more