[curriculum-eval] side-quest-10-01-agent-brief.md: score regression — largest decline in corpus (−0.84 from baseline 9.70)

[github-actions]

File: workshop/side-quest-10-01-agent-brief.md
Overall Score: 8.9 / 10.0 (corpus mean: 8.7)

Note: This file currently scores above corpus mean (8.9), but it shows the largest score decline in the corpus: −0.84 from a baseline of 9.70 over the 20-commit history window. This issue documents the regression and provides a recovery prompt.

Flagged Dimensions:

Dimension	Score	Benchmark	Delta vs. baseline
overall	8.86	9.70 (baseline)	−0.84

Root Cause (≤ 2 sentences):
The score history shows a sudden drop from 9.70 to 8.86 in the most recent commits, suggesting a content edit removed or weakened a previously strong dimension — most likely the scaffolding section, active-learning density, or checkpoint quality. The current metrics (scaffolding: 10, checkpoint: 10, active_learning: not flagged) indicate the drop may be driven by a concept-count or word-count increase in a recent edit.

Evidence (quoted from the file):

(Verify current file content — the rubric baseline at 9.70 indicates the prior version had higher scaffolding or active-learning scores than the current state reflects.)

Learning Science Rationale:
Score regressions in well-established workshop steps often trace to content expansion without corresponding activity additions — a pattern consistent with Split-Attention Effect (Sweller), where integrating new reference material into an existing activity page increases extraneous cognitive load without adding instructional benefit.

Improvement Prompt (for an agent):

Investigate workshop/side-quest-10-01-agent-brief.md to identify the regression source:
1. Run the curriculum rubric on the current file: python3 .github/skills/curriculum-quantitative-assessment/curriculum_assessment.py --file workshop/side-quest-10-01-agent-brief.md
2. Compare the output scores against the baseline (cognitive_load: expected ~10.0, active_learning: expected higher than current).
3. If word_count > 800 or new_concepts > 15 have increased since the baseline: trim the content that was recently added, or move new reference material to a dedicated side quest.
4. If active_learning density has dropped: add 2–3 short practice prompts with <details> reveal blocks.
5. Verify the checkpoint still has ≥ 4 specific checklist items.
Run: npx --yes markdownlint-cli2 "workshop/side-quest-10-01-agent-brief.md" to verify no lint errors.

Expected Score After Fix: 9.5 / 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 · 92.4 AIC · ⌖ 7.51 AIC · ⊞ 6.2K · ◷

• expires on Jul 29, 2026, 4:39 AM UTC