Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
63 lines (41 loc) · 3.41 KB

CONTRIBUTING.md

File metadata and controls

63 lines (41 loc) · 3.41 KB
Raw

Contributing to Agentic CRO (Growth-Pilot)

First of all, thank you for your interest in Agentic CRO! 🚀 This project envisions a "Fully Autonomous Frontend Optimization Machine." We believe a vision this immense requires the collective intelligence of the open-source community, not just a few lone engineers.

We welcome all forms of contributions: writing code, reporting bugs, proposing features, and expanding our multi-agent ecosystem. Please follow the guidelines below to ensure your Pull Requests (PRs) merge as smoothly as possible.

🛠 Open Architecture (Plug-and-Play MCP)

This project boasts a modular architecture entirely compatible with the MCP (Model Context Protocol) ecosystem.

1. Adding a New Connector Plugin (MCP Servers)

Do you want to integrate a new analytics tool (Amplitude, Mixpanel), experimentation platform (Optimizely, LaunchDarkly), or deployment service (GitLab/Bitbucket) alongside our existing posthog.ts, github.ts, and growthbook.ts?

You simply need to implement a connector interface inside our MCP servers directory.

// src/interfaces/YourAnalyticsInterface.ts
export interface IAnalyticsClient {
    connect(): Promise<void>;
    fetchDropoffData(funnelId: string): Promise<any>;
}

Implement this interface in a new file (e.g., src/mcp-servers/new-tool.ts), and submit a PR!

2. Adding a New VLM / QA Validator

Currently, our agents handle viewport capturing and VLM (Vision Language Model) visual inspection using Playwright. If you want to add an 'Accessibility (a11y) VLM Validator' or a 'Lighthouse Performance Validator', simply append your validation strategy and Temporal Activity inside the src/utils/ or src/orchestrator/ directories.

🌱 Are You New Here? (good first issue)

Due to the deep engineering footprint of this codebase (AST Mutation, Temporal Pipelines, Bayesian Thompson Sampling, Multi-Tenancy Architecture), contributing to the core engine might seem daunting. For first-time committers, we have established a strict good first issue policy:

  1. Check the Issues Tab: Look for issues tagged with good first issue in our GitHub repository.
  2. Examples of good first issues:
    • Enhancing context schema for agent prompts (src/agents/*.ts)
    • Adding missing jsdoc documentation to core temporal workflows
    • Minor CSS bug fixes in the UI mockup components (dashboard/)
    • Adding lightweight edge-case patches to the AST Parser

📝 Developer Environment Setup & PR Rules

  1. Fork & Branch: Fork this repository and create a local branch. (feature/your-cool-feature or fix/issue-name)
  2. Commit Convention: We strictly follow the Conventional Commits specification.
    • feat: Add new MCP connector for Amplitude
    • fix: Handle Promise rejection in GrowthBook fallback
    • docs: Update README Bayesian architecture diagram
  3. Dry-Run (Mandatory): Set IS_SHADOW_MODE=true in your .env to engage the local sandbox. Ensure your code does not spiral into an infinite loop by running at least one full cycle with the Temporal Worker.
  4. Pull Requests: Please attach proof of execution (e.g., screenshots of the running Agentic dashboard, or execution logs) in your PR description.

⚖️ Code of Conduct

We are much more wary of bad manners than bad code. Please maintain respect and empathy for your peers during code reviews.

We eagerly await your incredible PRs! Happy Hacking! 💻✨