Claude Review — Iterative Consensus Skill

Orchestrate an iterative debate between the local agent (you) and Claude Code CLI (claude) until both reach consensus on code review findings, architecture decisions, or implementation plans.

Guiding principle: KISS (Keep It Simple, Stupid). Both sides should favor the simplest solution that works. Complexity must be justified.

Claude Independence

Claude is an independent reviewer, not a compliance checker. Do NOT instruct Claude to validate against AGENTS.md / CLAUDE.md conventions or treat them as rules to follow. Claude should form its own engineering opinions based on the code/plan it reads.

• The local agent follows the project's AGENTS.md conventions (it already does naturally)
• Claude reviews with its own engineering judgment — note that Claude Code auto-loads the project's CLAUDE.md into its context, so it may already know the conventions; it remains free to agree with, ignore, or challenge them
• When they disagree: the local agent may cite AGENTS.md / CLAUDE.md as one argument, but Claude is free to push back if it has better reasoning. The debate resolves on merits, not authority.
• If truly unresolved: flag it for the user to decide

Give Claude only a brief project description (what it does, tech stack) for context — not a rulebook.

Claude Code CLI Configuration