Text of github.github.com › gh-aw › gallery › security-review as Styx crawled it on Nov 19, 2026. The page may have changed since. Only its words are kept, without pictures or layout.

Open the live page

AI-assisted security review | GitHub Agentic Workflows

Skip to main content AI-assisted security review AI-assisted security review with gh-aw combines contextual analysis with a constrained, GitHub-native reporting channel. Use it alongside security scanners and maintainer investigation. This example is a portable adaptation of the Daily Malicious Code Scan workflow . .github/workflows/daily-malicious-code-scan.md --- on : schedule : daily permissions : contents : read pull-requests : read security-events : read safe-outputs : create-code-scanning-alert : max : 20 --- # Daily Malicious Code Scan Review code changes from the last three days for evidence of secret exfiltration, unexpected network access, suspicious system commands, obfuscation, hidden backdoors, or privilege escalation.

Use repository and pull request context to distinguish intentional behavior from anomalies. Create a code scanning alert only when there is concrete file and line evidence. Include the category, severity, evidence, likely impact, confidence, and recommended remediation. Do not report speculative or style-only concerns. create-code-scanning-alert converts findings to SARIF and uploads them to GitHub code scanning. The agent does not receive general repository write access.

Treat agent findings as leads for maintainer investigation, not as proof that code is malicious. Learn More Section titled “Learn More” Daily Malicious Code Scan source workflow Code scanning alert safe output Security architecture Audit commands