commit-security-scan

You can scan code changes for security vulnerabilities by analyzing staged changes, single commits, or branch diffs to detect threats using LLM-powered reasoning and repository-specific threat model patterns before they reach production.

Yes, automated PR review maps identified vulnerabilities to CWE classifications and STRIDE categories, generating structured JSON reports that include severity assessments, exploit scenarios, and remediation guidance for each finding.

Yes, you can run a pre-commit security scan on uncommitted working directory changes, as the scanning scope supports staged changes, commit ranges, pull requests, and branch comparisons.

To detect SQL injection or missing authorization in feature branch commits, you can scan the commit range to identify vulnerabilities across all STRIDE categories using LLM reasoning and repository-specific threat models.

The best way to automate security checks across pull requests and branch diffs is to use LLM-powered reasoning that evaluates code changes and generates structured reports with severity levels, exploit scenarios, and remediation guidance.

No external dependencies are required to perform LLM-powered code review assistance, as the skill operates independently to analyze code changes and produce structured security findings with STRIDE classifications and CWE mappings.