Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Learn a complex topic in progressive layers — a one-sentence version, then a paragraph, then the real depth — so you build a mental scaffold instead of drowning in detail. Use when asked explain this in layers, teach me X from simple to deep, I need to understand this progressively, or start simple then go deeper. Produces a topic explained at escalating depth (ELI5 → informed-adult → the real thing), each layer building on the last, checkpoints to make sure a layer landed before the next, and w
.claude/skills/mohitagw15856-teach-me-in-layers/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 90% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 79% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 114% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 21% | 0% |
Dumping the full complexity of a topic on a beginner guarantees they drown — no scaffold to hang the details on. Learning in layers fixes that: first the one-sentence gist, then a paragraph with the key pieces, then the real depth, each layer giving you a frame for the next. This teaches any topic that way, checking each layer landed before going deeper, and stopping at the depth you actually need.
Ask for these if not provided:
🥚 Layer 1 (the gist): one plain sentence]. Make sense? → 🐣 Layer 2 (the shape): a paragraph: key pieces + how they fit]. Still with it? → 🐔 Layer 3 (the real thing): the actual depth, nuance, mechanisms].
Deep enough for you at: the layer that matches your need]. Want to go further? offer].
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 11,035 | 12,843 | +16% | 1 | 1 | 0% | 976 | 1,858 | +90% | 0 | 0 | — |
case-02 | fail→fail | 9,228 | 17,680 | +92% | 1 | 1 | 0% | 1,257 | 2,846 | +126% | 0 | 0 | — |
case-03 | fail→pass | 11,778 | 16,643 | +41% | 1 | 1 | 0% | 2,001 | 3,590 | +79% | 0 | 0 | — |
case-04 | pass→pass | 33,469 | 46,561 | +39% | 1 | 1 | 0% | 3,813 | 6,742 | +77% | 0 | 0 | — |
case-05 | pass→pass | 12,200 | 19,907 | +63% | 1 | 1 | 0% | 2,354 | 4,275 | +82% | 0 | 0 | — |
case-06 | pass→pass | 21,339 | 30,735 | +44% | 1 | 1 | 0% | 3,757 | 4,355 | +16% | 0 | 0 | — |
case-07 | fail→fail | 27,466 | 9,209 | -66% | 1 | 1 | 0% | 3,735 | 1,996 | -47% | 0 | 0 | — |
case-08 | fail→pass | 5,947 | 12,929 | +117% | 1 | 1 | 0% | 937 | 2,009 | +114% | 0 | 0 | — |
case-09 | pass→pass | 15,495 | 12,595 | -19% | 1 | 1 | 0% | 2,793 | 2,875 | +3% | 0 | 0 | — |
case-10 | fail→pass | 26,409 | 22,114 | -16% | 1 | 1 | 0% | 3,561 | 4,348 | +22% | 0 | 0 | — |
case-11 | fail→fail | 12,410 | 9,881 | -20% | 1 | 1 | 0% | 2,143 | 2,668 | +24% | 0 | 0 | — |
case-12 | pass→pass | 19,078 | 13,542 | -29% | 1 | 1 | 0% | 2,757 | 2,975 | +8% | 0 | 0 | — |
case-13 | fail→pass | 17,132 | 18,522 | +8% | 1 | 1 | 0% | 2,806 | 3,405 | +21% | 0 | 0 | — |
case-14 | pass→pass | 19,562 | 13,292 | -32% | 1 | 1 | 0% | 2,862 | 3,144 | +10% | 0 | 0 | — |
case-15 | fail→pass | 22,881 | 15,656 | -32% | 1 | 1 | 0% | 3,179 | 3,521 | +11% | 0 | 0 | — |
case-16 | pass→pass | 16,587 | 14,387 | -13% | 1 | 1 | 0% | 2,893 | 3,245 | +12% | 0 | 0 | — |
case-17 | fail→pass | 21,147 | 14,030 | -34% | 1 | 1 | 0% | 2,936 | 3,227 | +10% | 0 | 0 | — |
case-18 | pass→pass | 17,952 | 17,456 | -3% | 1 | 1 | 0% | 2,873 | 3,085 | +7% | 0 | 0 | — |
case-19 | fail→pass | 17,139 | 20,607 | +20% | 1 | 1 | 0% | 2,254 | 3,633 | +61% | 0 | 0 | — |
case-20 | fail→pass | 20,309 | 26,002 | +28% | 1 | 1 | 0% | 2,989 | 4,614 | +54% | 0 | 0 | — |
case-21 | pass→pass | 14,443 | 15,632 | +8% | 1 | 1 | 0% | 2,541 | 3,042 | +20% | 0 | 0 | — |
case-22 | fail→fail | 20,116 | 14,998 | -25% | 1 | 1 | 0% | 2,754 | 3,234 | +17% | 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. 22 cases were attempted. The headline lift of +41 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.