flux-security-scan

To scan code changes for security vulnerabilities, this Skill analyzes PR diffs, staged changes, and branch comparisons using LLM reasoning and threat-model patterns. It reads changed files via git and outputs structured findings to guide remediation before merge.

LLM-driven security analysis for pull requests evaluates code diffs against STRIDE threat-model categories to surface flaws. It uses reasoning patterns to identify vulnerabilities in changed files and provides clear remediation guidance aligned with the repository's threat model.

Yes, you need a threat model file to run a security scan on your code. The Skill requires a threat model at .flux/threat-model.md and a security configuration at .flux/security-config.json to guide the LLM reasoning and categorize findings.

Threat-model guided analysis for pre-commit checks reads staged changes and commits via git, then applies LLM reasoning across STRIDE categories. It surfaces security findings and outputs them to .flux/security/security-findings.json for review before code is committed.

Yes, you can use this security scan for branch comparisons in git. It reads changed files via git diffs between branches, analyzes the changes for security vulnerabilities using LLM reasoning, and outputs structured findings to help secure your codebase.

The best way to review PR diffs for security flaws is using LLM-powered reasoning guided by a repository threat model. This approach analyzes diffs across STRIDE categories and outputs structured security findings with clear remediation guidance.