Workstation Posture — Template

Module 2, Lesson 2.5 deliverable · save and freeze

How to use this template. Fill every blank with something specific and true. Vague answers defeat the whole purpose.

When done: type the finished version into workstation-posture.md in your course working folder. Then copy it unchanged into /capstone/workstation-posture-v1.md — that frozen copy is your second capstone artifact (matching my-first-loop-v1.md from Module 1).

If you prefer paper first: fill this sheet by hand, then transcribe. The writing-by-hand step is part of the point.

Part 1 — Cost posture (4 lines)

Line 01Monthly cost ceiling, per provider, in dollars.

E.g., Anthropic: $20/mo. OpenAI: $10/mo. Must match the hard cap in the console.

Line 02How I know I am approaching it.

Which email address receives the warning? Do I also check the usage dashboard on a schedule? When?

Line 03What I do when I hit it.

Stop-and-audit vs. raise-the-cap. Commit to one default. No reflex cap-raises.

Line 04High-volume rule.

Before running anything on a schedule or in batch, what small dry-run and paper estimate do I commit to?

Part 2 — Privacy posture (4 lines)

Line 01My default placement rule.

Cloud-default or local-default? One sentence, no hedges.

Line 02Categories that always go local.

At least three concrete examples — journal, family finances, medical, confidential drafts, etc.

Line 03Categories that are fine in cloud.

Schoolwork, public data, open-source code, writing you plan to publish, etc. Be specific.

Line 04The “when unsure” rule.

Default-to-local, or redact-and-use-cloud. One committed move, not a “trust my gut.”

Part 3 — Default-model rule (from Lesson 2.1)

One sentence: what’s your default (local / cloud / hybrid) and the specific condition that flips you to the other?

Part 4 — Installed stack

Local runner

e.g., Ollama + Llama 3.1 8B Instruct Q4

Cloud provider(s)

e.g., Anthropic primary, OpenAI secondary

Editor

e.g., VS Code 1.88

Terminal

e.g., macOS Terminal / Windows Terminal (PowerShell)

Coding agent(s)

Cowork, Claude Code, or both

Default coding agent & switch condition

e.g., Cowork by default; Claude Code when I want raw shell output

Key storage

e.g., .env in course folder, .env in .gitignore

Capstone folder location

e.g., ~/ai-architect-academy/capstone/

Part 5 — The four-part new-tool checklist (adopt, don’t rewrite)

Copy these four questions verbatim into your workstation-posture.md. You are adopting them, not editing them. Ask them every time you install a new tool.

  1. What data does this tool see, and where does it go? One sentence from the tool’s docs or settings. If I can’t find it, I don’t use it on real data yet.
  2. What does it cost, and how do I know I am approaching a limit? Same cap-and-warning pattern as my cost posture.
  3. What can it do on my behalf without asking? Read-only? Writes files? Sends messages? Moves money? The more autonomous, the more carefully I gate it.
  4. Where does it live in my posture? Local-only, cloud-default, or restricted to certain data categories?

Part 6 — Sign and date

Signed (student)
Date

Reminder — two files, not one

  1. workstation-posture.md in your course working folder (revisable; the living document).
  2. /capstone/workstation-posture-v1.md (frozen copy; the portfolio artifact).