How does Agent Triage work?

In 'AGENT_DECIDES' mode, the skill automatically implements high-priority errors and warnings while deferring low-value or stylistic suggestions unless otherwise specified.

Can I choose which comments the AI fixes?

Yes, by selecting 'HUMAN_SELECTS' mode, you can view a numbered list of comments and choose exactly which ones you want the agent to address.

What are the prerequisites for using this skill?

You must have the GitHub CLI (gh) installed and authenticated, and you need to set a PROPEL_API_KEY environment variable.

Does this skill work with non-Propel comments?

No, this skill is specifically designed to target Propel-authored comments and will ignore comments from other users or bots to maintain focus.

Why does it use a custom script instead of standard gh pr view?

The custom script uses the GitHub GraphQL API to access 'reviewThreads', which provides deeper access to inline comments that the standard CLI command often misses.

Propel PR Comment Handler

Name: Propel PR Comment Handler
Author: propel-gtm

bypropel-gtm

0•

Collaboration Tools

Automatically identifies and resolves Propel AI code review comments in GitHub Pull Requests using the GitHub CLI.

This skill streamlines the post-review workflow by fetching and triaging Propel AI-authored comments on GitHub Pull Requests. It utilizes the GitHub CLI and custom Python scripts to navigate review threads, filter out resolved discussions, and provide structured modes for implementation—ranging from full automation to human-guided selection. By categorizing issues by severity (error, warning, info), it helps developers prioritize critical fixes and maintain code quality with minimal manual overhead.

Key Features

010 GitHub stars

02Standardized execution plans and reporting for clear visibility into code changes.

03Automated fetching of inline PR comments and review threads via GitHub GraphQL API.

04Automatic severity inference (Error/Warning/Info) for better prioritization.

05Three execution modes: Full automation, AI-driven triage, or manual selection.

06Intelligent filtering of resolved threads and non-actionable summary comments.

Use Cases

01Automating the resolution of repetitive style suggestions or warnings from Propel AI.

02Rapidly addressing high-severity AI-detected bugs in a complex pull request.

03Guided triage of PR feedback where the developer wants to approve specific changes before application.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add propel-gtm/propel-code-skills propel-address-pr-comments

For use in Claude.ai and ChatGPT

Key Features

010 GitHub stars

02Standardized execution plans and reporting for clear visibility into code changes.

03Automated fetching of inline PR comments and review threads via GitHub GraphQL API.

04Automatic severity inference (Error/Warning/Info) for better prioritization.

05Three execution modes: Full automation, AI-driven triage, or manual selection.

06Intelligent filtering of resolved threads and non-actionable summary comments.

Use Cases

01Automating the resolution of repetitive style suggestions or warnings from Propel AI.

02Rapidly addressing high-severity AI-detected bugs in a complex pull request.

03Guided triage of PR feedback where the developer wants to approve specific changes before application.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add propel-gtm/propel-code-skills propel-address-pr-comments

For use in Claude.ai and ChatGPT