skills.rest
kotahashihama
kotahashihamaCOMMUNITY
Explore all skills by this author →
Website𝕏@kotahashihamaGitHub

loop-ai-reviews

Runs repeated AI code review rounds on GitHub pull requests until comments reach zero.

Updated Dec 25, 2021
Search Tags:#automation#workflow#code review#pull request#github#gh cli#ai review
One-click install
npx skills add https://github.com/kotahashihama/dotfiles --skill loop-ai-reviews-kotahashihama
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: loop-ai-reviews
Source: https://github.com/kotahashihama/dotfiles/tree/main/home/.claude/skills/loop-ai-reviews
Command: npx skills add https://github.com/kotahashihama/dotfiles --skill loop-ai-reviews-kotahashihama

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Manually triggering AI reviews on a pull request, waiting for the workflow, addressing feedback, and repeating the cycle is tedious and error-prone. This Skill automates the full loop: request an AI review, wait for the GitHub Actions workflow to finish, resolve any inline comments, and repeat up to a capped number of rounds. ## Core Features & Use Cases - Automated review loop: Orchestrates /request-ai-reviews, /wait-ci, and /resolve-ai-reviews as one continuous cycle per round. - Round control and cost guardrails: Accepts --rounds N (default 1, capped at 5), stops early when zero comments remain, and does not count failed or cancelled workflow runs as completed rounds. - Failure handling and finalization: Aborts after two consecutive workflow failures with run URL and conclusion, and invokes /finalize-pr once at the end if any fixes were pushed. - Use Case: You open a PR and say "run AI reviews 3 rounds" — the Skill requests the review, waits for the AI Review workflow, fixes the bot's inline comments, pushes the changes, re-reviews the new diff, and stops early if a round comes back clean. ## Quick Start Run AI reviews on the current pull request for up to 3 rounds and resolve any feedback automatically.

Frequently Asked Questions about loop-ai-reviews

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run multiple AI review rounds on a GitHub pull request?▼

Pass the PR number or URL with --rounds N, for example up to 5 rounds. Each round requests an AI review, waits for the AI Review workflow, resolves inline comments, and pushes fixes before starting the next round.

How do I automate fixing AI review comments on a PR?▼

The loop counts bot inline comments via the GitHub API after each workflow completes and invokes the resolve step to address and push fixes. Rounds with zero comments are skipped and stop the loop early.

What happens if the AI review workflow fails or is cancelled?▼

Failed or cancelled runs are not counted as completed rounds and the same round is retried. After two consecutive failures the loop aborts and reports the conclusion and run URL for investigation.

Is there a limit on how many review rounds I can run?▼

Yes, the maximum is 5 rounds and larger values are clamped to 5. The default is 1 round because each round incurs billing, and the loop also stops early once a round produces zero comments.

Can I force all rounds to run even when there are no review comments?▼

Yes, explicitly requesting that all N rounds run disables the early stop on zero comments. By default the loop terminates as soon as a round returns no feedback to avoid unnecessary billing.

Related Skills

Explore More Software Engineering Skills→
ConductionNLConductionNL
1

create-pr

Creates GitHub pull requests with branch validation, local checks, and conventional commit titles.

Official
Advanced
dimakisdimakis

land-pr

Shepherd GitHub pull requests through CI checks, reviews, and squash merges.

Community
Intermediate
erigontecherigontech
3.6k

erigon-ci

Runs Erigon CI test groups locally and dispatches GitHub Actions workflows on any branch.

Official
Intermediate
supersuit-techsupersuit-tech
4

review

Automate multi-round GitHub pull request reviews with inline comments.

Official
Advanced