Overview

10 projects shipped over 5 years. The Studio is the public brand layer; basecamp + terralabs + platform-ctl are the platform; triage + mlship + rxp + konfig + pulse are the OSS artifacts; ops-handbook is the operational journal. Plus services/llm-gateway/ and services/aiops/ living inside basecamp.

This index is the cast list for the program. Every project that ships in ROOT — whether a small Year 1 fluency CLI or the Year 5 capstone — has a row here, plus a PLAN.md you can read when its phase activates. The projects exist in service of the curriculum, not the other way around: each one earns its place either by deepening a pattern, by proving fluency in a tool, or by becoming a load-bearing piece of the platform itself.

The grouping below isn’t arbitrary. It’s the same A/B/C split as the Master Plan — Group A is the stage (the Studio + the substrate the rest runs on), Group B is the actors (services that run on the platform), Group C is the craft (operational discipline + the small fluency artifacts that prove you can ship code with PR review, CI, and releases). Read project plans when their phase activates, not before — they go stale fast if you read them out of context.

Why 10 projects, not 30

Most year-long curriculums ship a portfolio of disconnected demos. ROOT ships a small number of projects that each earn a real integration role inside basecamp. rxp feeds aiops log analysis; konfig validates Helm values in CI; pulse emits Prometheus metrics. Nothing is orphaned. That constraint is what makes the platform feel like one product instead of a collection of weekend hacks.

---

The 10 projects (three groups)

Group A: the Studio (the user-facing brand layer)

Project	Role	First touched
studio	Abukix Studio — the public brand. Portal + command palette + platform-ctl + 5 composition recipes layered on top of basecamp	Y5 P29 (launch)

Group B: the platform (the substrate)

Project	Role	First touched	Public when
basecamp	The GitOps repo of YAML that declares the 9-tier platform	Y1 P7	Y3 P19 launch
terralabs	IaC: Terraform modules + Crossplane Compositions	Y2 P9	Y2 P9 launch
platform-ctl	The unified CLI front-door	Y2 P12	Y5 P29 launch

Group C: services + craft (run on the platform; document operating it)

Project	Role	First touched	Public when
triage	First real service-on-K3s: on-call dashboard	Y1 P7	Y1 P7 quiet ship
mlship	Capstone OSS: one-command model deploy	Y4 P21 (v0)	Y5 P30 capstone launch
rxp	Regex CLI (Y1 Python fluency)	Y1 P4	Y1 P4 quiet ship
konfig	Config validator (Y1 Python fluency)	Y1 P4	Y1 P4 quiet ship
pulse	Probe scanner (Y1 Go fluency)	Y1 P5	Y1 P5 quiet ship
ops-handbook	Personal operational journal: runbooks, incidents, postmortems, ADRs, weekly logs	Y1 P1	Private; sanitized extract considered Y5

Inside basecamp (not standalone repos)

Sub-project	Role	First touched
basecamp/charts/services/llm-gateway/	Y4 flagship: OpenAI-compat LLM gateway with RAG, streaming, multi-model	Y4 P21 (v0) → P24 (v1) → P25 (v1.5)
basecamp/charts/services/aiops/	Y5: agent that operates the platform (incident triage, runbook execution, pattern detection)	Y5 P28
basecamp/charts/personal/personal-blog/	Personal: your blog hosted via basecamp	Y2
basecamp/charts/personal/personal-api/	Personal: life-data REST API (commits, weekly logs)	Y3 P18
basecamp/charts/personal/notes-rag/	Personal: RAG over your own writing	Y4 P24
basecamp/charts/personal/home-dash/	Personal: internal dashboard pulling from above	Y5 P29

---

Per-project structure

Each project gets its own directory with at least:

projects/{name}/
├── PLAN.md            ← what it is, scope by year, deliverables, success criteria
└── (later, when active)
    ├── BUILD-GUIDE.md ← how to build it (pattern-first; written when first active)
    └── LAUNCH.md      ← how to ship + market it (written near launch phase)

projects/studio/ additionally has composition.md — the 5 Studio composition recipes.

PLAN comes first, BUILD-GUIDE comes when the phase activates

Writing a build guide for a Y4 project in Year 1 is wasted effort — the tooling will have shifted, the patterns will be deeper, and your understanding of the trade-offs will be different. PLAN.md captures the durable parts (vision, scope, deliverables, success criteria); BUILD-GUIDE.md is written in-the-moment when the phase actually starts and the decisions are real.

---

Launch energy budget across 5 years

The discipline: launch energy is reserved for artifacts that earn it. Quiet ships push to GitHub + a tagged release; loud ships add blog post + LinkedIn + Show HN.

Year	Loud launches	Quiet ships
Y1	—	rxp, konfig, pulse, triage
Y2	terralabs	(basecamp + platform-ctl stay private)
Y3	basecamp (P19)	personal-api
Y4	— (llm-gateway v1 lives inside basecamp’s repo)	mlship v0/v0.5 stays private; notes-rag private
Y5	Abukix Studio + platform-ctl + aiops marketing (P29) + mlship v2 capstone (P30)	home-dash private

Loud launches per year averages ≤2. That’s the sustainable cadence; more dilutes attention.

---

Cross-cutting threads

• The platform substrate (basecamp + terralabs + platform-ctl + portal) grows continuously across all 5 years.
• The personal services (blog → personal-api → notes-rag → home-dash) prove the platform serves you, not just demos. One per year, Y2-Y5.
• The ops-handbook accumulates across all 5 years. By Y5 it’s ~140 runbooks, ~25 postmortems, ~250 weekly logs, 15+ ADRs.
• Upstream contributions: 1+ PR per phase = 30+ PRs across the program. Tracked in ops-handbook/contributions/.

---

Status (as of repo creation)

Started:        not yet — Phase 1 begins when Year 1 docs are reviewed
                          + homelab cleared
Public OSS:     0 (terralabs is the first, in Y2 P9)
Private:        all 10 projects scaffolded, no code yet