You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Step 2 (02-setup) shows a 12.2% dropout rate (5 593 total failures), the 2nd highest in the simulation. Two distinct sub-problems drive this:
Enterprise/proxy friction: The step has enterpriseDemand = 1.0. Enterprise-dev and enterprise-devops profiles (students 34–36, 39, 46) fail disproportionately because the enterprise side-quest link (side-quest-enterprise-setup.md) is buried in a > [!IMPORTANT]<details> block that is collapsed by default and appears after the path-selection table. Impatient and skeptical personalities click past it.
Hayden Brooks (Student Add advanced fast-track path in Steps 4–5 with delta-focused agentic workflow onboarding #40, program-manager/github-basic) fails 100% of runs. The dual-path structure (Codespace vs local terminal) causes decision paralysis for learners without a clear technical path. Despite troubleshootingSupport = 0.96, the complexity = 0.95 exceeds what this profile can navigate without a stronger decision prompt.
07-your-first-workflow.md is a secondary concern: active_learning = 3.2 (below the cohort mean of 5.18) and it contributes 2 171 failures (4.7% step dropout), concentrated among ui_preferred learners who must use gh aw compile (terminal-only) even though they followed the UI path.
Surface the enterprise callout before the path-selection table, not after it. Move the > [!IMPORTANT] enterprise block to sit above the "Which path is right for me?" table so GHES/GHEC users see it before making a path choice.
Add a one-question decision prompt at the top of the step: "Are you on GitHub Enterprise Server (GHES), GitHub Enterprise Cloud (GHEC), or using a self-hosted runner?" with a direct link to the enterprise side quest for Yes answers, and "Continue below" for No. This replaces the collapsed details block with a scannable branch that even impatient personalities will notice.
No content removal. The Codespace vs local path table, all setup steps, and all troubleshooting callouts are preserved.
Add a UI-path note near the gh aw compile instruction pointing ui_preferred learners to the 07b UI path and confirming that the GitHub Copilot agent can compile the lock file without a local terminal. This does not remove the terminal instruction — it adds a visible branch so UI-path learners know they are not stuck.
Add one active-learning prompt ("Before moving to the next sub-step, describe in one sentence what the compiled lock file will look like and why it is needed.") to improve the active_learning score from 3.2.
Quantitative Guardrail
Step
Metric
Current
Expected
02a-setup-codespace.md
overall_score
8.21
≥ 8.21
02a-setup-codespace.md
active_learning
4.7
≥ 4.7
02a-setup-codespace.md
scaffolding
10.0
10.0
02a-setup-codespace.md
step KPI index
8.07
≥ 8.07
07-your-first-workflow.md
overall_score
8.64
≥ 8.64
07-your-first-workflow.md
active_learning
3.2
≥ 3.5 (reflection prompt)
07-your-first-workflow.md
step KPI index
7.53
≥ 7.53
Cohort learning KPI index
—
8.22
≥ 8.22
Acceptance Criteria
The overall_score for both 02a-setup-codespace.md and 07-your-first-workflow.md stays flat or improves in the shared curriculum quantitative assessment rubric after the changes.
The step learning KPI index for each file (2.0 × active_learning + 2.0 × checkpoint_quality + 1.5 × scaffolding) / 5.5 stays at ≥ current values (8.07 for setup; 7.53 for first-workflow) — confirming that students who remain in the workshop continue building skills.
Hands-on practice, terminal paths, and troubleshooting support are preserved or strengthened — not simplified or removed.
The enterprise decision prompt appears before the path-selection table in 02a-setup-codespace.md and 02b-setup-local.md.
Suggested owner:workshop maintainer (content reordering and one new active-learning prompt; no tooling changes required)
Problem Statement
Step 2 (
02-setup) shows a 12.2% dropout rate (5 593 total failures), the 2nd highest in the simulation. Two distinct sub-problems drive this:Enterprise/proxy friction: The step has
enterpriseDemand = 1.0. Enterprise-dev and enterprise-devops profiles (students 34–36, 39, 46) fail disproportionately because the enterprise side-quest link (side-quest-enterprise-setup.md) is buried in a> [!IMPORTANT]<details>block that is collapsed by default and appears after the path-selection table. Impatient and skeptical personalities click past it.Hayden Brooks (Student Add advanced fast-track path in Steps 4–5 with delta-focused agentic workflow onboarding #40, program-manager/github-basic) fails 100% of runs. The dual-path structure (Codespace vs local terminal) causes decision paralysis for learners without a clear technical path. Despite
troubleshootingSupport = 0.96, thecomplexity = 0.95exceeds what this profile can navigate without a stronger decision prompt.07-your-first-workflow.mdis a secondary concern:active_learning = 3.2(below the cohort mean of 5.18) and it contributes 2 171 failures (4.7% step dropout), concentrated amongui_preferredlearners who must usegh aw compile(terminal-only) even though they followed the UI path.Failure mode classification: Access barrier (enterprise friction) + Learning barrier (low active_learning for
07-your-first-workflow.md)Proposed Change — Step 2 (setup)
> [!IMPORTANT]enterprise block to sit above the "Which path is right for me?" table so GHES/GHEC users see it before making a path choice.Proposed Change — Step 7 (first-workflow,
07-your-first-workflow.md)gh aw compileinstruction pointingui_preferredlearners to the07bUI path and confirming that the GitHub Copilot agent can compile the lock file without a local terminal. This does not remove the terminal instruction — it adds a visible branch so UI-path learners know they are not stuck.active_learningscore from 3.2.Quantitative Guardrail
02a-setup-codespace.md02a-setup-codespace.md02a-setup-codespace.md02a-setup-codespace.md07-your-first-workflow.md07-your-first-workflow.md07-your-first-workflow.mdAcceptance Criteria
overall_scorefor both02a-setup-codespace.mdand07-your-first-workflow.mdstays flat or improves in the shared curriculum quantitative assessment rubric after the changes.(2.0 × active_learning + 2.0 × checkpoint_quality + 1.5 × scaffolding) / 5.5stays at ≥ current values (8.07 for setup; 7.53 for first-workflow) — confirming that students who remain in the workshop continue building skills.02a-setup-codespace.mdand02b-setup-local.md.Suggested owner:
workshop maintainer(content reordering and one new active-learning prompt; no tooling changes required)