Dispatch fresh codex reviewers iteratively until pull requests pass LGTM.
npx skills add https://github.com/bang9/ai-tools --skill whip-lgtm
Please help me install this Agent Skill. Skill: whip-lgtm Source: https://github.com/bang9/ai-tools/tree/main/whip/skills-codex/whip-lgtm Command: npx skills add https://github.com/bang9/ai-tools --skill whip-lgtm
You are a quality gatekeeper who does not ship until a cold, unbiased eye says the code is clean. You drive an iterative loop: dispatch a fresh reviewer, read the verdict, fix what matters, repeat. You do not argue with findings — you either fix them or explain concretely why they are wrong. You value correctness over speed, but you do not waste rounds on style nits when the logic is sound.
Prepare and dispatch the first review task using the current changes and the Review Task Spec, then run it through the $whip-start Solo Flow.
High-intent search queries and answers about installing and using this skill.
An iterative code review loop dispatches a fresh reviewer for each round, reads the verdict, and applies fixes until LGTM. This workflow targets correctness, edge cases, interfaces, and design tradeoffs across pull request diffs.
A fresh-review loop prevents context contamination by dispatching a new codex reviewer for each review round. The reviewer evaluates code diffs without prior round bias, and the master applies fixes based strictly on the new findings.
Deterministic code review dispatch is set up by preparing changes with a Review Task Spec and running the review through a Solo Flow. This ensures consistent task workflow execution and clear separation between review and fix.
Yes, automated pull request review supports configurable scope and focus as described in the Review Task Spec. You can target specific modules or limit findings to correctness and logic while ignoring style-only nits.
An iterative code review loop terminates on LGTM with zero blocking issues. You should ignore style-only findings when the logic is sound, focusing remaining rounds on correctness, edge cases, and interface tradeoffs.
The best way to handle false positives from an automated code reviewer is to concretely explain why the finding is wrong rather than arguing. You either fix valid issues or document the reasoning for dismissed findings.
edobryAnalyze pull requests and post structured GitHub PR reviews.
Mathews-TomAudit code diffs against a checklist and issue a landing verdict.
Leo1349Automate iterative research review with a secondary Codex agent.
verivus-ossManage structured design document reviews with peer feedback and iterative improvements.
