Getting Started Guide

Relevant source files

The following files were used as context for generating this wiki page:

• README.md
• scripts/autofix.py
• .github/workflows/pr_agent.yml
• .pr_agent.toml
• scripts/utils.py (Integrated logic from autofix utility functions)

The pr-agent-settings repository centralizes configuration and automated repair tools for the PR-Agent across all repositories owned by @JWEB0689. This system enables global settings inheritance, repository-specific overrides, and autonomous code repair based on review findings or CI failures.

Sources: README.md:1-5, scripts/autofix.py:5-11

Configuration Management

PR-Agent uses a hierarchical configuration structure. The central .pr_agent.toml file in this repository provides global defaults for all child repositories.

Global Inheritance and Overrides

1. Global Settings: PR-Agent loads .pr_agent.toml from JWEB0689/pr-agent-settings for every pull request.
2. Local Overrides: You can define a .pr_agent.toml in individual repositories to override global settings.

Sources: README.md:7-14

Interactive Commands

You trigger PR-Agent features by commenting on a Pull Request.

Command	Action
/review	Performs code review, security, and test analysis.
/describe	Generates a PR description and changelog.
/improve	Provides commit-ready code suggestions.
/ask <question>	Answers specific questions about the changes.

Sources: README.md:20-25

Autonomous Repair System

The autofix.py script acts as an autonomous engineer. It fixes issues flagged by PR-Agent or repairs the codebase when CI builds fail.

Operational Modes

The script operates in two distinct modes depending on the PR state:

• Review-Fix Mode: Processes inline comments and review summaries left by PR-Agent. It sends findings to Gemini to generate updated file contents.
• CI-Fix Mode: Runs when no review findings exist but CI checks fail. It executes local test commands and attempts to repair the specific errors found in the logs.

Sources: scripts/autofix.py:7-13, scripts/autofix.py:284-307

Repair Workflow

The following diagram illustrates the decision logic and execution flow of the autonomous fixer:

The script prioritizes review findings before falling back to CI failure repairs.

Sources: scripts/autofix.py:284-315, scripts/autofix.py:228-260

Environment Configuration

The autonomous fixer requires specific environment variables to interact with the GitHub API and Gemini AI services.

Variable	Requirement	Description
GEMINI_API_KEY	Required	Google AI Studio key for Gemini models.
GEMINI_MODEL	Optional	Defaults to gemini-3.8-flash.
GH_TOKEN	Required	GitHub token for the gh CLI and REST API calls.
REPO	Required	The owner/repo string for the target project.
PR_NUMBER	Required	The ID of the pull request to process.
HEAD_REF	Required	The branch name for pushing repairs.

Sources: scripts/autofix.py:17-25, scripts/autofix.py:302-306

Integration Procedures

To integrate these settings into a new repository, follow these steps:

1. Configure Workflow: Reference the reusable workflow located at .github/workflows/pr_agent.yml in your repository's workflow directory.
2. Define Tests: Ensure your repository contains a supported manifest file (e.g., package.json, Cargo.toml, pyproject.toml) so the script can detect and run your test suite.
3. Set Secrets: Add GEMINI_API_KEY and GH_TOKEN to your repository secrets to enable autonomous repairs.
4. Trigger Repairs: Use the /autofix command (implied by script logic) or let the CI workflow trigger autofix.py on failure.

Sources: README.md:16-18, scripts/autofix.py:59-78

The central configuration system ensures consistent AI-assisted code reviews and self-healing CI pipelines across the @JWEB0689 organization by utilizing global defaults and autonomous repair scripts.