Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a realistic plan to learn a language for your actual goal — travel, conversation, work, or fluency — focused on what moves the needle instead of endless app streaks. Use when asked to help me learn [language], make a language learning plan, how do I get conversational, or study a language efficiently. Produces a goal-and-level read, a prioritized plan (the high-frequency vocab and core patterns first), a daily/weekly routine mixing input, speaking, and review, how to get real practice and
.claude/skills/mohitagw15856-language-learning-plan/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -2% | 0% |
People stall learning languages by doing the wrong things intensely — grinding app streaks and grammar drills while never speaking, or aiming vaguely at "fluency" with no plan. This builds a plan around your real goal, front-loads what actually gets you communicating (high-frequency vocabulary, core patterns, and speaking early), and sets a sustainable routine — with honest timelines instead of fantasy.
Ask for these if not provided:
Focus first: high-frequency vocab + core patterns for the goal]. Weekly routine
Milestones: goal-tied checkpoints]. Honest timeline: realistic for the goal] — consistency beats intensity.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 22,639 | 21,537 | -5% | 1 | 1 | 0% | 2,835 | 3,302 | +16% | 0 | 0 | — |
case-02 | pass→pass | 30,120 | 15,528 | -48% | 1 | 1 | 0% | 3,646 | 3,385 | -7% | 0 | 0 | — |
case-03 | fail→fail | 30,702 | 27,073 | -12% | 1 | 1 | 0% | 3,735 | 3,967 | +6% | 0 | 0 | — |
case-04 | pass→pass | 19,644 | 22,303 | +14% | 1 | 1 | 0% | 3,009 | 3,522 | +17% | 0 | 0 | — |
case-05 | pass→pass | 22,075 | 17,776 | -19% | 1 | 1 | 0% | 2,912 | 3,647 | +25% | 0 | 0 | — |
case-06 | fail→pass | 20,192 | 10,977 | -46% | 1 | 1 | 0% | 2,526 | 1,836 | -27% | 0 | 0 | — |
case-07 | fail→fail | 12,944 | 15,300 | +18% | 1 | 1 | 0% | 2,025 | 3,415 | +69% | 0 | 0 | — |
case-08 | fail→pass | 21,599 | 18,248 | -16% | 1 | 1 | 0% | 2,603 | 3,189 | +23% | 0 | 0 | — |
case-09 | fail→pass | 19,012 | 14,940 | -21% | 1 | 1 | 0% | 2,866 | 3,212 | +12% | 0 | 0 | — |
case-10 | pass→pass | 16,471 | 12,805 | -22% | 1 | 1 | 0% | 1,858 | 2,955 | +59% | 0 | 0 | — |
case-11 | pass→pass | 20,783 | 14,993 | -28% | 1 | 1 | 0% | 2,515 | 3,019 | +20% | 0 | 0 | — |
case-12 | pass→pass | 11,910 | 12,230 | +3% | 1 | 1 | 0% | 1,246 | 2,059 | +65% | 0 | 0 | — |
case-13 | pass→pass | 14,532 | 15,722 | +8% | 1 | 1 | 0% | 1,786 | 2,690 | +51% | 0 | 0 | — |
case-14 | pass→pass | 11,370 | 15,033 | +32% | 1 | 1 | 0% | 1,146 | 2,604 | +127% | 0 | 0 | — |
case-15 | pass→pass | 23,587 | 23,216 | -2% | 1 | 1 | 0% | 3,267 | 3,980 | +22% | 0 | 0 | — |
case-16 | fail→pass | 20,670 | 20,690 | +0% | 1 | 1 | 0% | 2,763 | 3,609 | +31% | 0 | 0 | — |
case-17 | pass→pass | 22,271 | 23,335 | +5% | 1 | 1 | 0% | 2,965 | 4,116 | +39% | 0 | 0 | — |
case-18 | pass→pass | 18,896 | 22,295 | +18% | 1 | 1 | 0% | 3,161 | 3,514 | +11% | 0 | 0 | — |
case-19 | pass→pass | 15,447 | 21,550 | +40% | 1 | 1 | 0% | 2,502 | 3,773 | +51% | 0 | 0 | — |
case-20 | pass→pass | 14,937 | 13,042 | -13% | 1 | 1 | 0% | 2,060 | 2,935 | +42% | 0 | 0 | — |
case-21 | fail→pass | 20,217 | 15,304 | -24% | 1 | 1 | 0% | 3,196 | 3,138 | -2% | 0 | 0 | — |
case-22 | fail→fail | 24,872 | 25,733 | +3% | 1 | 1 | 0% | 3,037 | 3,686 | +21% | 0 | 0 | — |
case-23 | fail→pass | 15,786 | 18,682 | +18% | 1 | 1 | 0% | 2,601 | 3,276 | +26% | 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 +26 percentage points is the difference between those two pass rates over the 23 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.