[workshop-sim] Repair: Split concept-overloaded 04-github-actions-intro.md into a focused refresher plus follow-on comparison

[github-actions]

Problem statement

04-github-actions-intro.md is the lowest-quality-scoring page in the entire curriculum (overall_score 5.39/10) and the second-highest-dropout Part 1 step (16.5% conditional dropout among 44,200 at-risk simulated runs, 95% MC interval 16.2%–16.9%), driven entirely by concept-overload (7,294 failures). The page packs a YAML refresher, a runner explainer, a five-row Classic-vs-Agentic comparison table, a label-the-parts activity, and a "try it" real-workflow exploration into one page with active_learning scoring only 3.9/10 despite containing several exercises — the exercises are dense and back-loaded rather than interleaved with the explanatory content.

Proposed change

Split the single page into two: keep 04-github-actions-intro.md focused on the core YAML refresher (trigger/jobs/steps + the annotated example + the "label a sample workflow" activity), and move the "Why This Matters for Agentic Workflows" comparison table plus the "Try it: Explore a real workflow" browser activity into a short follow-on page (or a clearly separated section with its own mid-page checkpoint) so no single page introduces both the classic-Actions mental model and the agentic-workflow comparison at once. Add one short check-in question after the annotated YAML example, before the comparison table, asking learners to predict what an agentic workflow replaces.

Failure mode classification

Learning barrier (conceptual overload from concentrating multiple concept layers in one page — not an access/tooling barrier, since this step requires no terminal or credentials per its own "Before You Start" section).

Quantitative guardrail

Current scores for 04-github-actions-intro.md:

• overall_score: 5.39/10
• Weakest rubric dimensions: checkpoint_quality 0.0/10 (rubric detection artifact — page uses :white_check_mark: shortcode, not literal ✅) and active_learning 3.9/10
• Learning KPI index: (2.0×3.9 + 2.0×0.0 + 1.5×5.0)/5.5 = 2.78/10

Splitting the page and adding one interleaved check-in reduces cognitive_load (currently 6.8/10) without cutting any existing activity, and increases effective active_learning density per page (more frequent, smaller practice moments rather than fewer large ones), which should raise both dimensions and the corresponding scores rather than lower them.

Acceptance criteria

• The shared curriculum quantitative assessment overall_score for the affected step(s) stays flat or improves after the change (baseline: 5.39/10 for the single combined page)
• The learning KPI index (2.0 × active_learning + 2.0 × checkpoint_quality + 1.5 × scaffolding) / 5.5 stays flat or improves across the split pages (baseline: 2.78/10)
• All existing exercises (label-the-parts activity, "try it" real-workflow exploration, and comparison table) are preserved or strengthened — none are removed to shorten the page
• workshop/README.md and workshop/00-welcome.md curriculum tables are updated if a new file is introduced by the split

Suggested owner

workshop maintainer
Related to #3152

Generated by 🔬 Workshop Student Simulator · copilot · auto · 271 AIC · ⌖ 27.2 AIC · ⊞ 15.2K · ◷

• expires on Sep 11, 2026, 3:18 AM UTC