You will apply scoped prompt techniques to generate AI-assisted refactored code that compiles successfully and is ready for peer review. Topics covered include the three pillars of scoped prompting — file selection, explicit constraints, and step-by-step change plans — how to distinguish modification targets from read-only context files, how to write verifiable constraints that protect function signatures and downstream dependencies, and how to sequence change plan steps so each produces an independently testable output. Through a vignette video, an instructional video, two readings, a coach dialogue, a practice assignment, and a knowledge check quiz, you will design and document a complete scoped prompt plan for a legacy module refactor. By the end of this module, you will be able to construct scoped prompts that produce bounded, reviewable AI-generated patches aligned with their team's codebase.

You will evaluate AI-generated code changes for security, privacy, and secret-management compliance before approving them for merge. Topics covered include why AI-generated code introduces specific security risks such as credential exposure, insecure defaults, and unsafe dependencies, how to apply a shift-left security review approach, and how to use OWASP Top 10 criteria, GitHub Secret Scanning, and the NIST SSDF as evaluation frameworks. Through an instructional video, two readings, a coach dialogue, a role play, a practice assignment, and a knowledge check quiz, you will apply a five-step evaluation procedure — scanning for secrets, applying OWASP criteria, assessing privacy implications, checking dependencies, and documenting findings — to a realistic AI-generated pull request. By the end of this module, you will be able to produce a written security assessment with a justified merge recommendation for any AI-generated contribution.