Route by consequence
GREEN work stays fast. Architecture, security, data and other consequential changes enter a RED design or code gate under explicit project policy.
AI CODE REVIEW GOVERNANCE
More code needs
more accountable review.
2nd Engineer gives engineering leaders a deterministic control system for AI-authored changes: which work receives independent review, what exact state was reviewed and who can approve the outcome.
Policy-routedExact-state evidenceHuman authorityReversible model roles
THE CONTROL GAP
AI review needs more than another opinion.
Separate author from reviewer
The authoring model does not grade its own work. A different model receives neutral design context or an immutable, read-only code snapshot.
Bind evidence to one state
Review evidence includes the change fingerprint, policy version and captured workspace state. New code makes the earlier result stale.
Keep authority human
The reviewer can pass, advise or block with evidence. Material disagreement ends at an explicit owner decision, not an automatic repair loop.
HOW 2ND ENGINEER WORKS
Assurance that follows the change.
2nd Engineer governs when a second model is called, what exact state it receives, what evidence it must return and who has authority to decide.
- 1
Classify locally
Deterministic project policy decides whether the work is routine or consequential without spending a model call.
- 2
Capture the review subject
2nd Engineer creates the neutral decision packet or secret-screened Git snapshot and calculates its fingerprint.
- 3
Obtain independent evidence
A separate reviewer returns structured findings, repository evidence, consequences and verification steps.
- 4
Record the accountable decision
The owner accepts, rejects with evidence or decides. Any later state change invalidates the earlier assurance.
AI CHANGE ASSURANCE
Give every consequential change an independent engineering checkpoint.
Start with one team, one policy and a reversible author-reviewer workflow.