[curriculum-eval] 05-agentic-workflows-intro.md: cognitive_load — 22 new concepts and low active learning density

[github-actions]

File: workshop/05-agentic-workflows-intro.md
Overall Score: 7.88 / 10.0 (corpus mean: 8.78)

Flagged Dimensions:

Dimension	Score	Benchmark	Delta
cognitive_load	6.5 / 10	≤ 15 new concepts (10)	−3.5
active_learning	2.9 / 10	≥ 3 activity density (10)	−7.1

Root Cause (≤ 2 sentences):
The step introduces 22 distinct new concepts (agentic workflow, task brief, frontmatter, lock file, trigger, permissions, runner, agent reasoning loop, tool use, sandbox, safe outputs, and more) across 1,154 words with zero code blocks, yielding an activity density of 0.87 — less than one third of the ideal. The heavy concept density with no hands-on code to anchor learning creates high intrinsic cognitive load for learners who are still building the schema for how GitHub Actions and AI agents interrelate.

Evidence (quoted from the file):

"An Agentic Workflow is a plain-English task brief that an AI agent executes inside GitHub Actions. You write what you want... and the agent reads your repo, calls tools, reasons about the results, and posts the output automatically. The frontmatter is fully Actions-compatible — triggers, permissions, and runners all apply."

Learning Science Rationale:
Sweller's Cognitive Load Theory identifies concept overload (>15 new elements per instructional unit) as a primary driver of extraneous cognitive load that crowds out schema formation. Mayer's Segmenting Principle recommends breaking complex explanations into learner-paced segments, ideally with worked examples or matching activities between segments. The existing "Try it: match the lifecycle parts" activity is a positive signal, but with 22 concepts competing for attention, three additional interactive checkpoints are needed to distribute the load.

Improvement Prompt (for an agent):

Edit workshop/05-agentic-workflows-intro.md to reduce cognitive load and increase active learning:
1. Audit the "Three things to know" and opening sections and split any list item that introduces more than one new term into two separate items. Aim to surface no more than 15 distinct new terms across the whole page.
2. Add at least one fenced code block (yaml or text) showing a minimal but complete agentic workflow file — frontmatter + task brief — with inline comments labeling each part.
3. After the existing "Try it: match the lifecycle parts" activity, add a second short activity: ask the learner to write (in one sentence) the difference between the .md source and the .lock.yml compiled file before revealing the answer in a <details> block.
4. Do not add numbered headings. Run markdownlint after editing.

Expected Score After Fix: 8.8 / 10.0

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• awmgmcpg

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "awmgmcpg"

See Network Configuration for more information.

Generated by 🔬 Curriculum Quality Evaluator · 86 AIC · ⌖ 7.39 AIC · ⊞ 6.2K · ◷

• expires on Jul 26, 2026, 4:39 AM UTC