Multi-tier code review automation
ralph-review-triopluginsetup L2★3,035
davepoon/buildwithclaude ↗What it does
Run three-tier sequential code review loop
Best for
Comprehensive solo code reviews where progressive model escalation catches breadth then depth
Inputs
- · Code diff or PR changes
Outputs
- · Surface-level issues (Haiku)
- · Logic/correctness review (Sonnet)
- · Deep architectural review (Opus)
Requires
- · Claude API (Haiku, Sonnet, Opus)
Preconditions
Code changes staged, valid Claude API keys, restart-on-fail enabled
Failure modes
API rate limits, context length exceeded, earlier tier crash blocks downstream, conflicting findings
Trust signals
- · Battle-tested on hundreds of issues
- · sequential tier escalation
- · restart-on-fail