『Why AI Coding Agents Leak Secrets in CI/CD (And How to Stop It)』のカバーアート

Why AI Coding Agents Leak Secrets in CI/CD (And How to Stop It)

Why AI Coding Agents Leak Secrets in CI/CD (And How to Stop It)

無料で聴く

ポッドキャストの詳細を見る

Your developer didn't leak an API secret to GitHub. Their autonomous AI coding agent generated a mock key, filled in a real one, and opened a Pull Request.

This episode breaks down the growing security risk of AI coding agents in CI/CD pipelines: why agents hardcode real credentials into test configs, why PR code reviews happen too late to prevent git history contamination, and how to enforce pre-receive hooks and ephemeral identity controls.

If your pipeline relies on prompts to prevent secret leaks, your security boundary is imaginary.

Keywords: AI coding agents, secret scanning, CI/CD security, DevSecOps, git push protection, credential leaks, GitHub Copilot agents, application security, software supply chain, OIDC tokens

This is Maya. New episodes three times a week.

youtube.com/@mayabuildsai

adbl_web_anon_alc_button_suppression_t1
まだレビューはありません