Safe PR Review Agent
Version updated for https://github.com/tushardhole/pr-review-agent to version v1.0.1.
- This action is used across all versions by ? repositories.
Action Type
This is a Node action using Node version 20.
Go to the GitHub Marketplace to find the latest changes.
Action Summary
The PR Review Agent GitHub Action leverages an AI-powered large language model (LLM) to automate code reviews for pull requests. It analyzes PR metadata, changed files, and repository conventions, generating inline review comments with severity tags and recommendations to ensure code quality and adherence to standards. This action streamlines the code review process by supporting configurable models, context-aware file analysis, and customizable review parameters, reducing manual effort and improving review efficiency.
What’s Changed
PR Review Agent v1.0.0
Initial stable release of tushardhole/pr-review-agent for GitHub Actions Marketplace.
Highlights
- AI-powered pull request review on
pull_requestevents - OpenAI-compatible provider support (
openai_base_url+model) including Groq/OpenAI/OpenRouter/Azure/Ollama - Structured review output with robust normalization for provider response variants
- Decision badges in top-level review comments (
✅/🚨/💬) while using safeCOMMENTevent posting - Inline comment safety:
- filters comments to resolvable diff anchors
- fallback retry to summary-only when GitHub cannot resolve a line
- Debug observability controls:
DEBUG_LLM_RESPONSEDEBUG_LLM_RESPONSE_MAX_CHARSDEBUG_LLM_RESPONSE_REDACT
- Strong test coverage (unit + integration) and strict TypeScript
- Release validation workflow for published releases
Action usage
- uses: tushardhole/pr-review-agent@v1
with:
github_token: ${{ secrets.GITHUB_TOKEN }}
openai_api_key: ${{ secrets.OPENAI_API_KEY }}
openai_base_url: https://api.groq.com/openai/v1
model: gpt-4o