The Developer Lifecycle

PlugZero is maintained with an Agent-First Philosophy. This means our workflows are designed not just for human engineers, but for a seamless collaborative loop between human creativity and AI-powered execution.

🚀 The Path to Productivity

Getting started on PlugZero is a tiered process designed to get you from “Clone” to “First Commit” in under 10 minutes.

Environments & Dual-Repo Orbit

You must have both the plugzero_api and plugzero-analytics repos cloned. They function as a “Binary Star System”—one cannot reach stable state without the other.

The Background Pulse (Redis & Celery)

Unlike simple web apps, PlugZero’s brain lives in the Background Queue.

You must have Redis running locally.
Without the Celery Worker, the SEO and Media crawlers will remain in a “PENDING” loop indefinitely.

The Identity Handshake

Once your backend is running, you must create a superuser and a test Team. The frontend relies on this “Identity Anchor” to render the dashboard layouts.

🧠 Collaborative Engineering (The Agent Loop)

We treat the Antigravity Agent as a senior pair programmer, not just a code generator.

The Documentation Duty

At PlugZero, a feature is not “Done” until it is narrative-documented.

The Living Doc Pattern: When the agent builds a new ViewSet or Component, it is trained to immediately update its corresponding page in this portal.
Reasoning over Registry: We prioritize explaining why a refactor was necessary over simply listing the lines of code changed.

Commit Hygiene

We follow the Conventional Commits standard strictly. This allows our internal tools to generate changelogs and release notes with zero human intervention.

feat(analysis): added isolation forest outlier logic
fix(ui): resolved radar chart clipping on mobile

🛠️ Testing Philosophy

We don’t believe in 100% test coverage for its own sake. Instead, we focus on “Intelligence Integrity.”

Logic Guards: Every computational function in processors.py must have an associated unit test ensuring the “Math stays true” even after a version update.
Visual Regressions: We focus on “State-Based” testing for the UI. Does the dashboard correctly transition from “Loading” to “Insight” across all 5 analysis types?