The Trust Lifecycle

From Idea to Installation

1

Submission

A user submits metadata (Name, Repo URL, Icon) via the Issue-Ops pipeline.

2

Verification

A human maintainer audits the Source Repository to ensure it is the official/original source.

3

Indexing

Bot scans Releases page. It extracts the direct download link for the APK.

4

Distribution

This link is saved to mirror.json on the Edge CDN.

5

Installation

The user's device downloads the file using the link from mirror.json.

6

Updates

Orion detects new release tags, updates the link, and triggers an update instantly.

1. The "Issue-Ops" Pipeline

Orion Store replaces the traditional "Developer Console" with a familiar tool: GitHub Issues. This lowers the barrier to entry while maintaining a transparent audit trail.

# Example Submission Issue
name: "Orion App"
repo: "https://github.com/developer/orion-app"
icon: "https://github.com/developer/orion-app/blob/main/icon.png?raw=true"
description: "A serverless app store."
category: "Tools"

Bots automatically parse this YAML to validate the repository, ensuring the structure is correct before a human ever looks at it.

2. The Human Firewall

Automation handles the syntax, but maintaining trust requires human intuition. Every app is audited by a maintainer for:

3. Mirror Generation (The O(1) Bot)

Once an app is labeled verified, the mirror_generator.py bot takes over. It performs a stateless operation:

  1. Scans the approved repository's /releases API.
  2. Finds the latest asset matching the APK pattern.
  3. Updates the global mirror.json file with the new version tag and download URL.
"We do not re-host your APKs. We simply point the user to exactly where you put them. This means zero hosting costs for us, and full control for you."

4. Global Distribution

The updated mirror.json is pushed to the gh-pages branch. GitHub's global CDN propagates this change worldwide in seconds.

When a user opens Orion Store, they fetch this simplified JSON manifest. Because it relies on static hosting, the "Store" can withstand millions of concurrent users without crashing or costing a dime.