Skip to content

Your first assessment

This tutorial takes you from nothing to a finished, interpreted AI readiness assessment. By the end you'll have a report committed under assessments/ and you'll understand what every part of it means.

It should take about fifteen minutes.

Before you start

You need one supported AI tool with an account, and a repository to assess. See Prerequisites if you're unsure. This tutorial uses GitHub Copilot CLI; the Claude Code path is identical in spirit.

We'll assess a real repository — open your terminal in the root of one you know well, so you can sanity-check the findings.

Step 1 — Install the plugin

Install and sign in to the Copilot CLI, then add this marketplace and install the plugin:

npm install -g @github/copilot
copilot

Inside the copilot session:

/plugin marketplace add techtalk/ai-readiness-assessment
/plugin install ai-readiness-assessment@techtalkai

Confirm it's there:

/skills list

You should see ai-readiness-assessment in the list.

Step 2 — Run the assessment

From the same session, with your repository as the working directory:

/ai-readiness-assess

The assessment runs in four visible movements:

  1. Discovery. It reports the habitat documents it found — things like CLAUDE.md, HARNESS.md, CI workflows, a specs/ directory — each with the path and the markers that confirmed the match. It also lists what's absent. Read this: absences matter as much as presences.
  2. Questions. It asks 3–5 clarifying questions, one at a time — the things the filesystem can't tell it, like how the team actually works with the agent. Answer honestly; an inflated answer just produces a confidently-wrong score.
  3. Assessment. It places all fourteen model dimensions, reads your cognitive level, and computes the Habitat Build Gap.
  4. Report. It writes assessments/YYYY-MM-DD-assessment.md and prints a short summary in the chat.

Step 3 — Read the result

Open assessments/YYYY-MM-DD-assessment.md. Work through it top to bottom:

  • Habitat Maturity Profile — your placement on all fourteen dimensions, each reported with the model's verb (e.g. Testing: L3 — Verifying). The headline Habitat Maturity Level is the rounded mean, with the weakest dimensions named as the ceiling.
  • Level Assessment — your cognitive level (L0–L5) and the one discipline holding you back.
  • Habitat Build Gap — the signed gap and its regime. This is the headline signal: are your habitat and your thinking in step?

See Read the Habitat Build Gap for how to act on each regime.

Step 4 — Act on one thing

The report ends with three recommendations and one TechTalk engagement suggestion — deliberately one, not a menu. Pick the single recommendation tied to your weakest dimension or discipline and do it this week. Re-run the assessment next month; the gap should move.

Optional — share it

Ask the assessment:

Render this as a shareable HTML page.

You'll get a single-file, print-friendly report you can hand to a lead or a stakeholder. See Render the HTML report.

Where to go next