Site reorganisation, 14 asset runbooks, webMCP server, and content audit

int-docs is reorganised into four clear sections. 14 operational runbooks cover every Phenom asset. A webMCP server exposes all content to agent sessions. Content audit stamps confirm accuracy.

What changed

The top nav now has six entries: What’s New | Tutorials | How-to | Docs | Reference | Blog.

Previously the site exposed four mostly-empty placeholder sections and hid the existing docs/ content tree from the nav entirely. The main navigability failures were:

  • docs/ (the richest content on the site) had no nav entry
  • whats-new/ had no nav entry
  • tutorials/ and explanation/ were empty shells with no child pages

The fix: added Docs and What’s New to the nav; removed the empty Explanation entry; populated Tutorials with two real walkthroughs.

Tutorials section

Two articles added:

14 Phenom asset runbooks (How-to section)

Every asset monitored by nest-ops.thephenom.app now has a dedicated runbook under How-to / Runbooks:

Asset Runbook
www.thephenom.app Runbook
nest.thephenom.app Runbook
dev-nest.thephenom.app Runbook
nest-ops.thephenom.app Runbook
drop.thephenom.app Runbook
chat.thephenom.app Runbook
api.thephenom.app Runbook
api-staging.thephenom.app Runbook
analytics.thephenom.app Runbook
webmail.thephenom.app Runbook
phenom-dev-postgres (RDS) Runbook
phenom-prod-postgres (RDS) Runbook
AWS SES (thephenom.app mail) Runbook
Honeypot sensor (3.230.47.213) Runbook

Each runbook follows a consistent structure: what it is, deployment chain, common operations, verification command, failure modes, and related links.

webMCP server

A Bun-based MCP server (mcp-server.ts, port 3300) now runs alongside the Hugo container. It exposes the full doc corpus — 238 docs across 58 sections, 14 runbooks — via five tools:

list_runbooks | get_runbook | search_runbooks | list_docs | get_doc

Query from any Claude Code session:

mcp2cli --mcp http://int-docs-mcp.matthewstevens.org/mcp --toon list-runbooks

See Agent Access: Query int-docs via MCP for the full usage guide.

Content audit stamps

All 14 runbooks and key reference pages now carry a provenance stamp (the green “Verified” banner at the top of each page). The stamp records the audit date, auditor, and the source used to verify the content — asset-registry.yaml, a live curl probe, or confirmed source code.

Pages not yet audited carry no stamp; a missing stamp is a clear signal that the content has not been reviewed against live infrastructure.


Commits

71d4f45 — site reorganisation + 14 asset runbooks + webMCP server

04da2dd — navbar fix + tutorials + audit stamps + What’s New entry

git log --oneline -5 on the main branch of Phenom-earth/phenom-earth-docs shows the full change set.