Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply constructivist learning theory to design instruction based on active knowledge construction, scaffolding, and the zone of proximal development. Use this skill when the user needs to design learner-centered instruction, apply Vygotsky's ZPD or scaffolding principles, or evaluate learning environments for constructivist alignment — even if they say 'how do people really learn', 'student-centered design', or 'scaffolding for learners'.
.claude/skills/asgard-ai-platform-grad-constructivism/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -12% | 0% |
Constructivism holds that learners actively construct knowledge through experience and social interaction rather than passively receiving it. Rooted in Piaget (cognitive constructivism) and Vygotsky (social constructivism), it emphasizes that learning is situated, social, and built on prior knowledge.
Trigger conditions:
When NOT to use:
IRON LAW: Learning Occurs in the Zone of Proximal Development
Instruction too far ABOVE current level is ineffective (frustration).
Instruction AT current level produces no growth (boredom).
Effective learning happens in the ZPD — the gap between what a learner
can do alone and what they can do with guidance:
1. Identify current developmental level
2. Provide scaffolding (temporary support)
3. Gradually remove scaffolding as competence grows (fading)Identify what learners already know. Constructivism insists new knowledge is built ON existing schemas — without this assessment, instruction may miss or contradict prior understanding.
Create learning tasks situated in realistic contexts. Knowledge constructed in abstract, decontextualized settings transfers poorly.
Provide graduated support: modeling → coaching → prompting → fading. Match scaffolding intensity to the gap between current ability and task demands.
Facilitate collaborative learning, peer discussion, and negotiation of meaning. Knowledge is socially constructed through dialogue and shared activity.
markdown# Constructivist Learning Design: {Topic/Course} ## Learner Analysis - Prior knowledge: {what learners already know} - Developmental level: {current capability} - ZPD target: {what they can achieve with support} ## Learning Environment - Authentic task: {real-world situated problem} - Social interaction: {collaborative structures} - Multiple perspectives: {how diverse viewpoints are included} ## Scaffolding Plan | Phase | Support Level | Activities | Fading Strategy | |-------|-------------|------------|-----------------| | Initial | High | ... | ... | | Developing | Medium | ... | ... | | Advanced | Low | ... | ... | ## Assessment - Process assessment: {how learning construction is evaluated} - Authentic assessment: {performance in realistic contexts}
references/zpd-scaffolding.mdreferences/authentic-assessment.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 107,715 | 20,420 | -81% | 1 | 1 | 0% | 7,295 | 3,589 | -51% | 0 | 0 | — |
case-02 | fail→pass | 48,664 | 14,250 | -71% | 1 | 1 | 0% | 6,484 | 2,961 | -54% | 0 | 0 | — |
case-03 | fail→pass | 38,539 | 22,928 | -41% | 1 | 1 | 0% | 3,881 | 3,862 | -0% | 0 | 0 | — |
case-04 | fail→pass | 27,960 | 15,718 | -44% | 1 | 1 | 0% | 3,504 | 3,271 | -7% | 0 | 0 | — |
case-05 | fail→pass | 30,141 | 12,402 | -59% | 1 | 1 | 0% | 3,052 | 2,684 | -12% | 0 | 0 | — |
case-06 | pass→fail | 15,697 | 7,224 | -54% | 1 | 1 | 0% | 2,181 | 1,868 | -14% | 0 | 0 | — |
case-07 | pass→pass | 13,169 | 7,262 | -45% | 1 | 1 | 0% | 2,123 | 1,863 | -12% | 0 | 0 | — |
case-08 | pass→pass | 14,775 | 9,123 | -38% | 1 | 1 | 0% | 2,465 | 2,211 | -10% | 0 | 0 | — |
case-09 | pass→pass | 15,727 | 14,773 | -6% | 1 | 1 | 0% | 2,366 | 2,995 | +27% | 0 | 0 | — |
case-10 | pass→pass | 22,124 | 15,512 | -30% | 1 | 1 | 0% | 3,115 | 2,878 | -8% | 0 | 0 | — |
case-11 | pass→pass | 18,529 | 70,895 | +283% | 1 | 1 | 0% | 2,341 | 2,658 | +14% | 0 | 0 | — |
case-16 | pass→pass | 12,679 | 12,508 | -1% | 1 | 1 | 0% | 1,905 | 1,931 | +1% | 0 | 0 | — |
case-12 | pass→pass | 15,397 | 16,661 | +8% | 1 | 1 | 0% | 2,561 | 2,864 | +12% | 0 | 0 | — |
case-13 | pass→pass | 18,596 | 13,173 | -29% | 1 | 1 | 0% | 2,529 | 2,595 | +3% | 0 | 0 | — |
case-14 | fail→pass | 14,535 | 12,749 | -12% | 1 | 1 | 0% | 2,291 | 2,717 | +19% | 0 | 0 | — |
case-15 | pass→pass | 19,383 | 17,905 | -8% | 1 | 1 | 0% | 2,480 | 3,137 | +26% | 0 | 0 | — |
case-17 | pass→pass | 21,402 | 14,513 | -32% | 1 | 1 | 0% | 2,793 | 2,705 | -3% | 0 | 0 | — |
case-18 | pass→pass | 21,566 | 16,448 | -24% | 1 | 1 | 0% | 2,621 | 2,738 | +4% | 0 | 0 | — |
case-19 | pass→pass | 24,576 | 14,765 | -40% | 1 | 1 | 0% | 3,117 | 3,097 | -1% | 0 | 0 | — |
case-20 | pass→pass | 24,567 | 10,009 | -59% | 1 | 1 | 0% | 2,097 | 2,055 | -2% | 0 | 0 | — |
case-21 | pass→pass | 9,800 | 9,709 | -1% | 1 | 1 | 0% | 1,414 | 1,992 | +41% | 0 | 0 | — |
case-22 | pass→pass | 14,349 | 12,841 | -11% | 1 | 1 | 0% | 1,920 | 2,500 | +30% | 0 | 0 | — |
case-23 | fail→fail | 53,786 | 16,053 | -70% | 1 | 1 | 0% | 5,346 | 3,241 | -39% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +22 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.