The art of secure software: Zero-trust delivery pipelines
summary
Argues that CI/CD must be treated as a high‑risk asset and gives concrete controls: run CI runners as production, enforce MFA, centralize secrets in AWS Secrets Manager or HashiCorp Vault, automate dependency checks and SBOMs, and use CodeQL and Semgrep for continuous static analysis. Recommends continuous dependency updates to avoid red fatigue, run tabletop exercises, and use a preprod threat‑hunting sandbox with an AWS GuardDuty fitness function that blocks unexpected outbound callbacks.
AI-written summary · about
chapters (22)
- 0:00Introduction
- 1:48Secure delivery pipeline
- 2:31Zero trust meaning
- 4:20Zero trust principles
- 5:43Threat modeling CI/CD
- 8:08Network security in CI/CD
- 10:15IaaS vs SaaS for CI/CD
- 12:44User access and MFA
- 14:06Endpoint security attacks
- 15:30Dependency vulnerabilities
- 17:57Automating dependency checks
- 20:49Code vulnerabilities analysis
- 23:18Human factors in coding
- 26:43Developer least privilege
- 28:50Assume breach monitoring
- 31:58Threat hunting in CI/CD
- 33:02Incident response planning
- 34:05Managing secrets
- 35:30Environment isolation
- 37:58CI/CD runners security
- 39:05Threat hunting sandbox
- 41:12Summary
related talks
38mSecuring your CI/CD Pipeline
37mA Secure Continuous Delivery Pipeline
40mBest Practices for Securing CI/CD Pipelines or How to Get Security Right
25mBuilding a DevOps Supply Chain on Zero Trust Principles
48mImplement ZeroTrust with Dedicated DevSecOps Pipeline
48mCI/CD Security
28mMitigating OWASP Top 10 CI/CD with AWS
26mYour code might be secure, but what about your pipeline? ...