review

Automating PR review involves analyzing the diff against the base branch to surface SQL injection risks, LLM prompt leakage, and trust boundary violations. This proactively detects structural issues and unintended side effects, reducing manual review time and merge risk.

LLM trust boundary analysis evaluates code diffs to identify prompt leakage and unauthorized LLM access paths before code merges. It detects violations where untrusted inputs might bypass safety controls, ensuring prompt integrity and preventing unintended model behaviors in production.

Yes, automated diff analysis detects conditional side effects and SQL safety issues by evaluating code changes against the base branch. It surfaces potential SQL injection vectors and evaluates structural logic to identify unintended consequences before landing code.

Run pre-landing checks by executing the review skill directly on a PR diff to score risk factors. This analyzes the changes against the base branch, calculates automated risk scoring for SQL and LLM vulnerabilities, and provides proactive guidance for safe merges.

Automated PR diff analysis focuses on SQL safety, LLM trust boundaries, and structural side effects but may not catch complex business logic flaws or runtime vulnerabilities outside the diff context. It serves as a proactive pre-check to guide reviewers, not a replacement for comprehensive security audits.