File: workshop/05-agentic-workflows-intro.md
Overall Score: 5.43 / 10.0 (corpus mean: 6.15)
Flagged Dimensions:
| Dimension |
Score |
Benchmark |
Delta |
| cognitive_load |
8.5 |
≤800 words |
1,107 words (+307 over) |
| active_learning |
2.4 |
density ≥3 |
0.72 (2 code blocks + 6 checklist items / 1,107 words) |
| checkpoint_quality |
0.0* |
≥4 items |
*tooling false-negative (see corpus scorecard) |
Root Cause (≤ 2 sentences):
This page is 1,107 words — 307 over the 800-word cognitive-load ideal — and, despite containing three well-designed reveal-based activities ("Activity 1," "Activity 2: agentic or standard?," "Activity 3: write a task brief"), the automated activity-density metric only counts fenced code blocks (2) and checklist items (6), so the five <details>/reveal interactions that make up the page's actual pedagogical core are invisible to the density score, understating how interactive the page already is while its raw length still drags cognitive_load down.
Evidence (quoted from the file):
Task A: Run lint and unit tests on every pull request, fail if any check exits non-zero.
Reveal Task A answer
**Standard Actions workflow.** Every run follows the same fixed steps. No judgment required.
Learning Science Rationale:
Mayer's segmenting principle recommends breaking longer instructional content into learner-paced segments; this page already does that reasonably well via its <details> reveals, but at 1,107 words of surrounding prose (three key terms table, two-file-structure explanation, three separate diagrams) it still exceeds the point where Sweller's CLT predicts working-memory overload for a single continuous page, especially for a "core path" step near the beginning of the workshop where learners have the least prior schema to offload new terms onto.
Improvement Prompt (for an agent):
Open workshop/05-agentic-workflows-intro.md. At 1,107 words (ideal ≤800), this
early core-path page front-loads three diagrams, a terms table, and a two-file
explanation before its three well-designed reveal activities. Trim length
without removing the interactive `<details>` activities, which are the
strongest part of the page:
1. Shorten the "The two-file structure" section: the two bullet points
describing the `.md` source file and `.lock.yml` compiled file repeat
information already shown in the accompanying diagram's alt text. Cut the
bullets to one sentence each, keeping the "GitHub Actions runs this file,
not the `.md`" distinction since it's tested in the Checkpoint.
2. In "Three key terms," remove the sentence "For a full glossary, see Side
Quest: Agentic Workflows Deep Dive" and instead add that link inline next
to the table caption—this is a 1-line trim, not a content change.
3. Combine "Activity 2: agentic or standard?" intro paragraph with the
preceding diagram's caption instead of repeating similar framing language
twice.
4. Do not remove any of the three `<details>` activities, the diagrams, or the
Checkpoint section — they are the page's load-bearing interactive content.
5. Target a final word count of 850-900 words.
6. Run `npx --yes markdownlint-cli2 "**/*.md"` after editing and fix any lint
errors before finishing.
Expected Score After Fix: 6.2 / 10.0
Generated by 🔬 Curriculum Quality Evaluator · copilot · auto · 142.5 AIC · ⌖ 18.6 AIC · ⊞ 9.5K · ◷
File:
workshop/05-agentic-workflows-intro.mdOverall Score:
5.43 / 10.0(corpus mean:6.15)Flagged Dimensions:
Root Cause (≤ 2 sentences):
This page is 1,107 words — 307 over the 800-word cognitive-load ideal — and, despite containing three well-designed reveal-based activities ("Activity 1," "Activity 2: agentic or standard?," "Activity 3: write a task brief"), the automated activity-density metric only counts fenced code blocks (2) and checklist items (6), so the five
<details>/reveal interactions that make up the page's actual pedagogical core are invisible to the density score, understating how interactive the page already is while its raw length still drags cognitive_load down.Evidence (quoted from the file):
Learning Science Rationale:
Mayer's segmenting principle recommends breaking longer instructional content into learner-paced segments; this page already does that reasonably well via its
<details>reveals, but at 1,107 words of surrounding prose (three key terms table, two-file-structure explanation, three separate diagrams) it still exceeds the point where Sweller's CLT predicts working-memory overload for a single continuous page, especially for a "core path" step near the beginning of the workshop where learners have the least prior schema to offload new terms onto.Improvement Prompt (for an agent):
Expected Score After Fix:
6.2 / 10.0