critic

To identify gaps in a software implementation plan, use an unbiased reviewer to cross-check plan claims against actual source code. This process reads the plan, past lessons, and codebase to find integration mismatches, producing a detailed critique report with actionable suggestions.

A code-plan cross-check verifies whether your implementation plan's claims match the actual codebase. You need this unbiased review during the codebase planning phase to catch integration issues and hidden gaps before any code is actually written.

A fresh AI session improves code review by avoiding bias from previous context. It conducts a multi-round critique that incorporates lessons learned on the first round, updates the cost ledger, and uses the Opus model to thoroughly analyze source files and plan documentation.

Yes, you can use the Opus model to review codebase planning and detect integration issues. It reads your actual source files and implementation plans to generate a structured critique report, ultimately returning a clear pass or revise verdict.

The limitation of using an automated critic for risk analysis is that it requires reading the plan, past lessons, and actual source code to function properly. It is strictly designed for the codebase planning phase to provide a fresh critique, not for live code execution or runtime debugging.