File: workshop/04-github-actions-intro.md
Overall Score: 5.39 / 10.0 (corpus mean: 6.14)
Flagged Dimensions:
| Dimension |
Score |
Benchmark |
Delta |
| cognitive_load |
6.8 |
≤ 15 new concepts |
new_concepts=28 (concept_overload, −6.5 vs ideal) |
| active_learning |
3.9 |
density ≥ 3 |
activity_density=1.19 (4 code blocks + 5 checklist items / 7.56×100 words) |
| checkpoint_quality |
0.0 |
— |
tooling false negative (see companion regex-fix issue); page has a valid checkpoint with 5 items |
Root Cause (≤ 2 sentences):
The page introduces 28 distinct new terms/concepts (trigger, job, step, runner, action, on:, jobs:, runs-on:, self-hosted runner, agentic workflow, task brief, safe outputs, and more) in one pass across four separate YAML examples, which exceeds the 15-concept cognitive-load ceiling by nearly double even though the activities (labeling exercise, "explore a real workflow" task) are well-designed.
Evidence (quoted from the file):
Annotated example — each comment names the key term (this is a standard Actions workflow, not an agentic workflow):
on: workflow_dispatch # trigger: the event that starts this workflow
jobs:
hello: # job: a named group of steps on one machine
runs-on: ubuntu-latest # runner: the machine GitHub provisions for this job
Learning Science Rationale:
Sweller's Cognitive Load Theory predicts that introducing more than ~15 new schema elements in a single instructional segment overwhelms working memory (typically limited to 4±1 chunks), especially when the page also previews agentic-workflow vocabulary (task brief, safe outputs) that belongs to the next step rather than this one — this front-loads extraneous concepts before the learner has consolidated the Actions primitives being taught here.
Improvement Prompt (for an agent):
Edit workshop/04-github-actions-intro.md to reduce concept load:
1. Move the "Why This Matters for Agentic Workflows" comparison table (which introduces trigger/runner/steps/output mapped to agentic-workflow terms like "task brief" and "safe outputs") into workshop/05-agentic-workflows-intro.md instead, since those terms are the subject of that step, not this one. Replace it here with a single one-sentence forward-reference: "The next step, What Are Agentic Workflows?, builds on these same triggers and runners."
2. In the "Label a sample workflow" section, drop the `uses: actions/checkout@v4` action-vs-step distinction from this page (defer "action" as a term to a later steps-and-jobs deep dive) and keep the labeling exercise to the four core terms: trigger, job, runner, step.
3. Increase activity_density by converting the "Try it: Explore a real workflow" numbered steps into a checklist (`- [ ]`) so learners can mark progress, which also contributes toward the active_learning benchmark.
Keep the existing checkpoint section unchanged.
Expected Score After Fix: 6.5 / 10.0
Generated by 🔬 Curriculum Quality Evaluator · copilot · auto · 77.1 AIC · ⌖ 14.6 AIC · ⊞ 9K · ◷
File:
workshop/04-github-actions-intro.mdOverall Score:
5.39 / 10.0(corpus mean:6.14)Flagged Dimensions:
Root Cause (≤ 2 sentences):
The page introduces 28 distinct new terms/concepts (trigger, job, step, runner, action,
on:,jobs:,runs-on:, self-hosted runner, agentic workflow, task brief, safe outputs, and more) in one pass across four separate YAML examples, which exceeds the 15-concept cognitive-load ceiling by nearly double even though the activities (labeling exercise, "explore a real workflow" task) are well-designed.Evidence (quoted from the file):
Learning Science Rationale:
Sweller's Cognitive Load Theory predicts that introducing more than ~15 new schema elements in a single instructional segment overwhelms working memory (typically limited to 4±1 chunks), especially when the page also previews agentic-workflow vocabulary (task brief, safe outputs) that belongs to the next step rather than this one — this front-loads extraneous concepts before the learner has consolidated the Actions primitives being taught here.
Improvement Prompt (for an agent):
Expected Score After Fix:
6.5 / 10.0