In The Loop Labs Pair Review
in-the-loop-labs-pair-review is a Claude Code plugin for the Ship phase that runs AI-assisted code review with required human-in-the-loop feedback and implement-review-fix loops.
by in-the-loop-labs · github.com/in-the-loop-labs/pair-review
Run AI-assisted code review with explicit human-in-the-loop feedback loops before you merge PRs or ship fixes.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install in-the-loop-labs-pair-review@in-the-loop-labs/pair-reviewBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:in-the-loop-labs/pair-review") and the paste-ready config with skillselion.get_install_config("plugin:in-the-loop-labs/pair-review").
What it does
in-the-loop-labs-pair-review is a Claude Code plugin family built around one idea: AI can accelerate code review, but a human must stay in control of what ships. Solo builders who merge their own PRs get structured critic loops—analysis of diffs, explicit feedback, and implement-review-fix cycles—instead of trusting a single-pass model summary. The bundle includes two plugins and optional MCP server wiring, so you can run reviews locally, against a PR, or integrated with your agent toolchain. Keywords in the catalog point to pair programming, code-critic, and multi-level loops, which maps well to Ship-phase review habits: catch regressions, naming issues, and missing tests before customers see them. It is heavier than a lint rule and narrower than a full CI platform; think of it as a review copilot that waits for your OK. Register it in Claude Code when you are ready to treat review as a repeatable session, not a one-off prompt.
Highlights
- Human-in-the-loop AI code review—not fully autonomous merge bots.
- Includes MCP server integration plus standalone/local review modes (pluginCount: 2).
- Code-critic and feedback-loop patterns: analyze changes, critique, implement-review-fix cycles.
- Pair-programming and pair-review flows for pull requests and in-editor sessions.
- Three-level loop structure so you stay author of record while the agent drafts fixes.
Why builders use it
Solo builders ship faster with agents but still merge blind spots because one-shot AI review misses issues and lacks accountable fix cycles.
After install, you get structured pair-review loops where the agent critiques changes, proposes fixes, and you approve each step before merge.
At a glance
- Type - Plugin in Development Tools.
- Adoption - 0 installs, 27 stars, 0 votes.
FAQ
Who is in-the-loop-labs-pair-review for?
It is for Claude Code users—especially solo builders—who want AI help reviewing diffs and PRs but need to stay in the loop for every fix batch.
When should I use in-the-loop-labs-pair-review?
Use it before merging feature branches, during pair-programming on risky changes, and whenever you want implement-review-fix cycles instead of a single review comment.
How do I add in-the-loop-labs-pair-review to my agent?
Install both plugins from in-the-loop-labs/pair-review into Claude Code, configure the MCP server if you use server-side review, then start a pair-review or critic-loop session on your branch or PR.
Comments
Share how you use in-the-loop-labs-pair-review, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.