Detection of insecure dependencies
Dependabot performs a scan of the default branch of your repository to detect insecure dependencies, and sends Dependabot alerts when:
* A new advisory is added to the GitHub Advisory Database.
* The dependency graph for a repository changes. For example, when a contributor pushes a commit to change the packages or versions it depends on, or when the code of one of the dependencies changes.
Additionally, GitHub can review any dependencies added, updated, or removed in a pull request made against the default branch of a repository, and flag any changes that would reduce the security of your project. This allows you to spot and deal with vulnerable dependencies before, rather than after, they reach your codebase.
Note: When you push a commit to GitHub that changes or adds a supported manifest or lock file to the default branch, the dependency graph is automatically updated. In addition, the graph is updated when anyone pushes a change to the repository of one of your dependencies.

Excluding folders and files from secret scanning
You can customize secret scanning to automatically close alerts for secrets found in specific directories or files by configuring a secret_scanning.yml file in your repository.
Under Edit new file, type paths-ignore: followed by the paths you want to exclude from secret scanning. This tells secret scanning to automatically close alerts for everything in the docs directory.

When a supported secret is leaked, GitHub generates a secret scanning alert. Alerts are reported on the Security tab of repositories on GitHub, where you can view, evaluate, and resolve them.

Resolving code scanning alerts
From the security view, you can view, fix, or dismiss alerts for potential vulnerabilities or errors in your project's code.
Who can use this feature?
Users with write access
Note: Fixing an alert manually
Anyone with write permission for a repository can fix an alert by committing a correction to the code. If the repository has code scanning scheduled to run on pull requests, it's best to raise a pull request with your correction. This will trigger code scanning analysis of the changes and test that your fix doesn't introduce any new problems.