File: workshop/side-quest-11-06-anthropic-key.md
Overall Score: 5.39 / 10.0 (corpus mean: 6.14)
Flagged Dimensions:
| Dimension |
Score |
Benchmark |
Delta |
| cognitive_load |
7.2 |
≤ 15 new concepts |
new_concepts=26 (concept_overload, −5.5 vs ideal) |
| active_learning |
3.5 |
density ≥ 3 |
activity_density=1.05 |
| checkpoint_quality |
0.0 |
— |
tooling false negative (see companion regex-fix issue); page has a valid checkpoint with 5 items |
Root Cause (≤ 2 sentences):
This short (665-word) side quest crams 26 distinct concepts — API key generation, secret naming rules, whitespace pitfalls, engine frontmatter fields, network allow-lists, billing considerations — into a linear procedure with three separate <details> "common mistakes" call-outs, producing high concept density per word relative to the 15-concept ideal. The three nested > [!IMPORTANT] / > [!NOTE] / > [!TIP] callouts each introduce their own sub-concepts (key-visibility risk, billing risk, naming risk) rather than consolidating them.
Evidence (quoted from the file):
- Wrong name: any variation (
anthropic_api_key, ANTHROPIC-API-KEY, CLAUDE_API_KEY) will cause a silent auth failure. The name must be exactly ANTHROPIC_API_KEY.
- Copied with extra whitespace: pasting from some tools adds a leading space. Delete and re-create the secret if you are unsure.
- Closed the Anthropic tab before saving: you cannot retrieve the key again. Delete the key at [console.anthropic.com]((console.anthropic.com/redacted) and generate a new one.
- Network allow-list missing: the
claude engine needs outbound access to api.anthropic.com.
Learning Science Rationale:
Sweller's Cognitive Load Theory holds that split-attention and multiple simultaneous warnings (four distinct "common mistakes" bullet points plus three separate callout boxes) increase extraneous load without adding germane learning value, since a learner following the linear procedure only needs to see the mistake that is relevant to the step they are currently on, not all four at once before starting.
Improvement Prompt (for an agent):
Edit workshop/side-quest-11-06-anthropic-key.md to reduce concept density:
1. Move the "Common mistakes with this secret" <details> block's four bullet points inline next to the specific step they relate to, instead of grouping all four together after step 6. For example, place the "wrong name" and "whitespace" warnings directly under step 4 (setting the secret name/value), and the "closed the tab" warning directly under the key-generation section where it is already partially covered by the [!IMPORTANT] callout (merge them to avoid duplication).
2. Consolidate the three top-level callouts (IMPORTANT: key shown once, NOTE: billing, TIP: secret naming) into at most two by merging the billing NOTE into the surrounding pricing paragraph as plain prose rather than a callout, since it is supplementary context rather than an action the learner must take right now.
3. Keep the existing checkpoint section unchanged.
Expected Score After Fix: 6.3 / 10.0
Generated by 🔬 Curriculum Quality Evaluator · copilot · auto · 77.1 AIC · ⌖ 14.6 AIC · ⊞ 9K · ◷
File:
workshop/side-quest-11-06-anthropic-key.mdOverall Score:
5.39 / 10.0(corpus mean:6.14)Flagged Dimensions:
Root Cause (≤ 2 sentences):
This short (665-word) side quest crams 26 distinct concepts — API key generation, secret naming rules, whitespace pitfalls, engine frontmatter fields, network allow-lists, billing considerations — into a linear procedure with three separate
<details>"common mistakes" call-outs, producing high concept density per word relative to the 15-concept ideal. The three nested> [!IMPORTANT]/> [!NOTE]/> [!TIP]callouts each introduce their own sub-concepts (key-visibility risk, billing risk, naming risk) rather than consolidating them.Evidence (quoted from the file):
Learning Science Rationale:
Sweller's Cognitive Load Theory holds that split-attention and multiple simultaneous warnings (four distinct "common mistakes" bullet points plus three separate callout boxes) increase extraneous load without adding germane learning value, since a learner following the linear procedure only needs to see the mistake that is relevant to the step they are currently on, not all four at once before starting.
Improvement Prompt (for an agent):
Expected Score After Fix:
6.3 / 10.0