---
title: "Website Update Policy"
status: draft
audience: "AI agents and human maintainers of this website"
offer_type: page
primary_cta: none
last_reviewed: 2026-07-17
approval_required: true
ai_os_update_policy: propose-then-approve
related_services: []
---

# Approval-first update policy

This website is maintained under the same propose-then-approve discipline
Operon applies inside engagements. The rules:

1. **AI can propose updates.** Agents may draft changes to copy, offers,
   design, data files, and structure — as proposals, not publications.
2. **The human owner approves public claims.** Joshua (owner) must explicitly
   approve anything public-facing before it ships: claims, pricing, offers,
   case studies, proof-of-work entries, and policy changes.
3. **All approved changes are logged.** Every shipped change gets an entry in
   /CHANGELOG.md — what changed, when, and why.
4. **Rollback is always possible.** The site is version-controlled and content
   lives in discrete files under /content/ and /data/, so any change can be
   reverted cleanly.

## Hard content rules (apply to every proposal)

- No invented clients, testimonials, case studies, results, certifications,
  or partnerships.
- No pricing on the site until pricing is approved and published in
  /data/pricing.yaml.
- The Proof of Work section stays a placeholder until real, approved work
  ships.
- Tone stays practical and non-hype (see /DESIGN.md do/don't rules).

## For agents

Machine-readable site map and policies: /data/site-manifest.json.
Summaries: /llms.txt (short), /llms-full.txt (full).