Can it distinguish between bot and human comments?

Yes, it specifically looks for feedback from both human reviewers and automated tools like Sentry, Codecov, and Bugbot to ensure all requirements are met.

Does this skill require special permissions?

Yes, it requires the GitHub CLI (gh) to be installed and authenticated with permissions to view PRs, read checks, and access repository logs.

What happens if there is no PR for my branch?

The skill will identify that no PR exists for the current head and inform the user to create one before it can begin the iteration process.

Will it fix every CI failure automatically?

It targets identifiable code issues but is designed to stop and ask for help if a failure persists after 3 attempts or if the issue is related to infrastructure.

PR Iteration Agent

Name: PR Iteration Agent
Author: getsentry

bygetsentry

•

115

Deployment & DevOps

Automates the pull request lifecycle by fixing CI failures and addressing review feedback until all checks pass.

About

The Iterate PR skill streamlines the development workflow by managing the repetitive cycle of code submission and feedback. It monitors CI/CD pipelines via the GitHub CLI, extracts logs from failed builds, processes both human and bot review comments, and applies targeted fixes to the codebase. By automatically waiting for builds to complete and re-evaluating the PR status, it ensures that developers can focus on complex tasks while the skill handles the tedious feedback-fix-push-wait loop required to reach a green state.

Key Features

Consolidated feedback gathering from human reviewers and automated bots
Intelligent 'watch' functionality that monitors CI status in real-time
Validation logic to prevent redundant fixes and skip invalid feedback
115 GitHub stars
Context-aware code modifications to address specific linter and test failures
Automated CI failure log extraction and analysis via GitHub CLI

Use Cases

Managing the push-and-wait cycle for complex CI pipelines without manual polling
Fixing straightforward CI failures or linting errors across multiple files
Addressing a large batch of minor code review comments automatically

About

Key Features

Consolidated feedback gathering from human reviewers and automated bots
Intelligent 'watch' functionality that monitors CI status in real-time
Validation logic to prevent redundant fixes and skip invalid feedback
115 GitHub stars
Context-aware code modifications to address specific linter and test failures
Automated CI failure log extraction and analysis via GitHub CLI

Use Cases

Managing the push-and-wait cycle for complex CI pipelines without manual polling
Fixing straightforward CI failures or linting errors across multiple files
Addressing a large batch of minor code review comments automatically