pullfrog/pullfrog
Open-source model-agnostic BYOK GitHub bot that runs in GitHub Actions
What it solves
Pullfrog is a GitHub bot that integrates coding agents into the GitHub workflow. It eliminates the need to manually move context between a local IDE and GitHub, allowing developers to trigger AI agents to write code, review pull requests, and manage issues directly within the platform.
How it works
Pullfrog operates as a GitHub Action. It uses an internal MCP server to pull relevant repository context and can be triggered in three ways:
- Mention-based: Tagging
@pullfrogin a comment. - Dashboard-based: Triggering tasks via a web console.
- Event-based: Configuring automated triggers for events like issue creation, PR creation, or PR reviews.
It supports multiple LLM providers (Anthropic, OpenAI, Gemini, xAI, DeepSeek, OpenRouter, Moonshot) and can be used as a standalone step in custom GitHub workflows to generate structured JSON output via Zod schemas.
Who it’s for
Developers and maintainers who want to automate repository management and coding tasks using AI agents without leaving GitHub.
Highlights
- Automated Coding: Can implement features, create PRs, and automatically fix CI failures by reading logs.
- Automated PR Reviews: Configurable triggers to auto-review new pull requests.
- Automated Issue Management: Responds to common questions and creates implementation plans for new issues.
- Structured Output: Supports JSON schema validation for reliable integration into subsequent workflow steps.
- Agent Agnostic: Compatible with various major LLM providers.
Related
- Project
- Project
- Project
- Project
- Project