Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Interview your future self about a decision or a stuck moment — a structured perspective-shift that pulls you out of present emotion and into the long view, using your own values and patterns rather than generic advice or woo. Use when someone says 'I don't know what to do', 'help me think long-term about this', 'what would future me say', or is stuck in the fog of a big decision. Produces an interview with your 5-or-10-years-older self, the themes it surfaces, and one concrete next step it poin
.claude/skills/mohitagw15856-future-self-interview/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 92% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 71% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 160% | 0% |
Present-you is a bad advisor on big decisions — flooded with immediate emotion, sunk cost, and the loudest current pressure. Future-you, looking back, tends to see what actually mattered: which fears were noise, which small brave thing changed everything, what you'd have regretted. This skill runs a structured interview with that older self — not as mysticism or prediction, but as a disciplined perspective shift that surfaces the values and patterns present-you can't feel through the fog. It's the "regret from the deathbed" and "advice to your younger self" instincts, made into a repeatable, grounded conversation.
chosen life stage) answers — built from the user's own stated values, history, and patterns, not generic wisdom
dismiss, the thing they wish present-you had done sooner
future-you is honest, not a hype-man
without an action is just a nice feeling
Ask for (if not already provided):
raw material — future-you is built from these, not invented)
view for a life-shape question)
the user's own values and history — if the user has always regretted playing it safe, future-you knows that; if they've been burned by impulsiveness, future-you remembers. Generic "follow your heart" future-selves are useless; a future-you made of the user's real patterns is uncannily useful.
it?", "what did I waste time worrying about?", "what do you wish I'd done this year?", "what mattered that I can't see yet?" — and future-you answers specifically. The Q&A structure does the perspective work; a lecture doesn't.
longer view, which sometimes says "yes you should make the scary change" and sometimes "the thing you're agonizing over won't matter; the thing you're ignoring will." Future-you can deliver an uncomfortable truth; a purely soothing one has failed the exercise.
present-you's wishful thinking wearing a costume vs a genuine values-consistent long view. The honest version notes: "this answer might be your fear/hope talking — here's how to tell."
interview ends by asking future-you: "what's the one thing you wish present-me would do this week?" — small, concrete, doable now.
## The setup
[The decision · the horizon · who future-you is, built from your values/patterns]
## The interview
Present-you: [real question]
Future-you: [specific, honest, in-your-own-values answer]
[…several exchanges — worth it? / what I wasted worry on / what mattered / the regret]
## What future-you kept returning to
[The themes · the fears they dismissed · the thing to do sooner]
## Insight vs wishful thinking
[Where this is genuine long-view · where it might be hope/fear in disguise, and how
to tell]
## The one thing to do this week
[Small, concrete, now]perspective-shift tool; future-you is a lens, not an oracle
helps no one
future-you to a predetermined answer, name that gently
life-premortem]] is the shadow version (imagine it failed); regret-minimizer]] and decision-journal]] neighbors; the-time-capsule]] to actually write to your future self; franklin-decision-ledger]] for the analytical complement.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 36,418 | 26,023 | -29% | 1 | 1 | 0% | 3,013 | 4,118 | +37% | 0 | 0 | — |
case-02 | fail→fail | 21,670 | 26,353 | +22% | 1 | 1 | 0% | 2,929 | 3,954 | +35% | 0 | 0 | — |
case-03 | fail→fail | 16,252 | 15,213 | -6% | 1 | 1 | 0% | 2,142 | 2,726 | +27% | 0 | 0 | — |
case-04 | pass→pass | 41,315 | 23,531 | -43% | 1 | 1 | 0% | 3,332 | 3,573 | +7% | 0 | 0 | — |
case-05 | pass→pass | 21,073 | 20,072 | -5% | 1 | 1 | 0% | 2,217 | 3,406 | +54% | 0 | 0 | — |
case-06 | fail→pass | 46,073 | 24,323 | -47% | 1 | 1 | 0% | 3,609 | 4,703 | +30% | 0 | 0 | — |
case-07 | fail→pass | 21,701 | 19,770 | -9% | 1 | 1 | 0% | 2,206 | 4,226 | +92% | 0 | 0 | — |
case-08 | fail→fail | 17,739 | 15,704 | -11% | 1 | 1 | 0% | 2,486 | 2,811 | +13% | 0 | 0 | — |
case-09 | fail→pass | 22,695 | 23,334 | +3% | 1 | 1 | 0% | 2,319 | 3,976 | +71% | 0 | 0 | — |
case-10 | pass→pass | 23,805 | 18,855 | -21% | 1 | 1 | 0% | 2,879 | 4,023 | +40% | 0 | 0 | — |
case-11 | pass→pass | 19,185 | 23,645 | +23% | 1 | 1 | 0% | 2,105 | 3,991 | +90% | 0 | 0 | — |
case-12 | pass→pass | 19,872 | 20,540 | +3% | 1 | 1 | 0% | 2,372 | 3,681 | +55% | 0 | 0 | — |
case-13 | pass→pass | 20,979 | 25,291 | +21% | 1 | 1 | 0% | 2,358 | 4,210 | +79% | 0 | 0 | — |
case-14 | fail→pass | 14,793 | 19,409 | +31% | 1 | 1 | 0% | 1,413 | 3,674 | +160% | 0 | 0 | — |
case-15 | pass→pass | 24,402 | 22,650 | -7% | 1 | 1 | 0% | 2,832 | 3,759 | +33% | 0 | 0 | — |
case-16 | pass→pass | 16,369 | 25,861 | +58% | 1 | 1 | 0% | 2,574 | 4,208 | +63% | 0 | 0 | — |
case-17 | pass→pass | 23,684 | 16,577 | -30% | 1 | 1 | 0% | 2,670 | 3,723 | +39% | 0 | 0 | — |
case-18 | fail→pass | 18,174 | 28,032 | +54% | 1 | 1 | 0% | 2,712 | 3,648 | +35% | 0 | 0 | — |
case-19 | pass→pass | 21,331 | 21,003 | -2% | 1 | 1 | 0% | 2,425 | 3,716 | +53% | 0 | 0 | — |
case-20 | fail→pass | 13,410 | 14,660 | +9% | 1 | 1 | 0% | 2,169 | 3,687 | +70% | 0 | 0 | — |
case-21 | pass→pass | 25,183 | 20,068 | -20% | 1 | 1 | 0% | 3,820 | 4,070 | +7% | 0 | 0 | — |
case-22 | pass→fail | 21,739 | 27,989 | +29% | 1 | 1 | 0% | 3,327 | 5,076 | +53% | 0 | 0 | — |
case-23 | pass→fail | 20,612 | 30,055 | +46% | 1 | 1 | 0% | 2,023 | 5,782 | +186% | 0 | 0 | — |
case-24 | pass→fail | 19,236 | 20,389 | +6% | 1 | 1 | 0% | 2,198 | 3,983 | +81% | 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. 24 cases were attempted. The headline lift of +17 percentage points is the difference between those two pass rates over the 24 comparable cases. 3 cases got worse with the skill loaded, and they are 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.