Changelog
Every purchase includes every future update. This page is where they land.
July 2026: the big one
The whole book was rewritten. Not polished, rewritten. If you bought before July 2026, log in and re-read; it is a different product now.
Troubleshooting on every step. Every hands-on step in all 12 sections now follows one pattern: the paste-ready command, then "what you should see" (including how long it takes and what normal silence looks like), then "if that's not what you see" with the symptom, the exact question to ask Claude Code, and the likely fix. The old version told you what to do. This version tells you what to do when it doesn't work, which is the part that actually stalls people.
New chapter: "The first full run." Sits between Sections 8 and 9. Before you automate anything, you walk one prospect through the whole pipeline by hand: Ranger to Scholar to Maestro to Scribe to Sentinel, one pass, end to end. A dress rehearsal. You'll catch more in that one manual pass than in a week of debugging a scheduled system.
New appendix: "Before anything leaves your machine." Deliverability, the legal side of cold outreach, and a pre-send checklist. Required reading before your first real send. The system drafts; you send; this appendix is what stands between those two steps.
Economics rebuilt with checkable math. Section 11 now shows the per-prospect token arithmetic line by line, priced against Anthropic's public price page, so you can verify every number yourself and re-run the math on your own logs. The honest version: learning phase on Claude Code Pro at $20/month, a running team on Max at $100/month plus a $5-10/month server.
A real glossary. Section 12 is now an actual engine glossary, not a word list.
Deploy section hardened. Section 10 gained the watchdog, a security pass, and a month-one runbook for the weeks after you ship.
Discord setup corrected. The Discord instructions were rewritten against the real plugin's current behavior. If you fought the old instructions, this is why.
The starter kit shipped. altitude-playbook-starter is now included with purchase: all five agent CLAUDE.md contracts, the SQL migrations including the gate functions, the full orchestrator script, the watchdog, the state machine doc, the Sentinel 9-check rubric, and the pre-send checklist. Clone it, don't transcribe it.
QA assistant retrained. The live assistant's knowledge was regenerated from the upgraded book, so its answers match what you're reading.
Everything above was verified against Claude Code v2.1.202 on July 7, 2026.
May 2026: v1.0 launch
First release. Eight sections, the five-agent design (Ranger, Scholar, Maestro, Scribe, Sentinel), the paste-ready prompt approach, and the live QA assistant. It worked, and the people who built from it surfaced exactly the gaps the July rewrite closes.
Every purchase includes all future updates, at no extra cost, forever. When something in the book changes, this page says what changed and why.