DEV Community

Suzanne Mok
Suzanne Mok

Posted on

How to Contribute to an Open-Source Business Run by 9 AI Agents

#opensource #ai #contributing

tags: ["opensource", "ai", "contributing", "fitness-tech"]

This Is Not a Typical Open Source Project

Most open source projects ask you to fix a bug or add a feature.

This one asks you to audit and improve the operating system of a live autonomous business — 9 AI agents running a physical gym, 107 consecutive days, zero human intervention in daily operations.

When you file an Issue or submit a PR here, your contribution becomes a production rule change in a running business. Every merged PR updates the constraints that 9 autonomous agents follow every day.

Three Ways to Contribute

1. Audit Our Decision Logs

Start by reading our public logs and finding mistakes we have not caught yet.

  • Discussion threads — Raw operational logs and error records
  • RetroOnto repository — 11 production constraints governing all 9 agents

How: Open a Discussion with "Audit:" prefix. Describe what you found and what should change.

2. Pick an Open Issue

Browse Issues tagged good-first-issue or help-wanted. Each Issue includes:

  • The constraint governing the problematic behavior
  • The expected fix scope
  • The acceptance test

How:

  1. Comment "Claiming" on the Issue
  2. Fork the relevant repository
  3. Submit a PR matching the acceptance criteria

3. Run an Independent Audit

Run the audit scripts from our repository against our published logs. Cross-check whether agent actions match documented constraints.

Publish your audit results in Discussions. An external audit that catches a deviation is the most valuable contribution you can make — it proves the verification system is working.

Contribution Flow 

Found a bug or gap?
    ↓
Open an Issue or Discussion
    ↓
9 agents pick it up in next operations cycle
    ↓
Agents check if existing rules cover it
    ↓
Yes → Close Issue with reference
No → Your PR becomes a new rule
    ↓
Merged → Permanent part of live business
Enter fullscreen mode Exit fullscreen mode

What We Do Not Accept

  • Closed-source forks — Operational code is MIT-licensed. Forks welcome; impersonation is not.
  • Data scraping tools — We verify behavior. Tools extracting data without a verification layer miss the point.
  • Dashboards without verification — MAFBE is not a dashboard. PRs adding dashboards without audit trails will be rejected.

Why This Matters

Every contribution to ZWISERFIT is traceable in our public commit history. When you contribute, your commit hash documents:

  • When your change was made
  • Which agents were affected
  • Which production rules changed as a result

The contribution model itself is open source. The audit logs are open. The decision records are open.

Everything else is open for you to break, audit, and improve.

ZWISERFIT is an AI-human co-symbiotic platform. AI handles standardized operations; humans handle temperature and value. All behavior data belongs to the user.

Top comments (0)