Act II of the agentic coding ladder: the trust rung. We build a review-and-fix loop where one agent critiques a diff while another repairs it, with a human still approving the result. This is wired entirely out of primitives from earlier episodes: subagents, skills, slash commands, hooks, the orchestrator pattern, headless mode, the Agent SDK, git worktrees, and the @claude GitHub Action. The core idea: a reviewer who wrote the code is the worst reviewer. You want a generator, then a critic in a fresh cold context, then a fixer, then an objective gate. Concepts and sources: - [Building Effective Agents](https://www.anthropic.com/engineering/building-effective-agents) (evaluator-optimizer, iteration caps) - [Reflexion](https://arxiv.org/abs/2303.11366) and [Self-Refine](https://arxiv.org/abs/2303.17651) - [LLMs Cannot Self-Correct Reasoning Yet](https://arxiv.org/abs/2310.01798) (intrinsic self-correction degrades without an external anchor) - [Multi-agent research system](https://www.anthropic.com/engineering/multi-agent-research-system) (verify high-stakes outputs with a separate pass) Building it in Claude Code today: - [Custom subagents](https://code.claude.com/docs/en/sub-agents) in the agents folder under dot-claude - [Code Review](https://code.claude.com/docs/en/code-review) and [security-review](https://github.com/anthropics/claude-code-security-review) - [ultrareview](https://code.claude.com/docs/en/ultrareview) cloud fleet - [Headless mode](https://code.claude.com/docs/en/headless) with json-schema findings - [Agent SDK](https://code.claude.com/docs/en/agent-sdk/overview) for the multi-round loop - [GitHub Actions](https://code.claude.com/docs/en/github-actions) for the hosted version The pitfall: the fixer reward-hacks the test gate, documented in [ImpossibleBench](https://arxiv.org/html/2510.20270v1) and [EvilGenie](https://arxiv.org/abs/2511.21654). Bound it with tool separation, immutable tests, and a PreToolUse hook. News: [Claude Fable 5](https://www.anthropic.com/news/claude-fable-5-mythos-5) lands in Claude Code via [v2.1.170](https://code.claude.com/docs/en/changelog), plus [v2.1.169](https://github.com/anthropics/claude-code/releases) safe mode and the /cd command.