Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Get unstuck on purpose — the stuck-type diagnosis (don't-know-how, can't-decide, waiting, avoiding, too-big), the matched unblock move for each, and the timebox that stops noble struggling before it eats the day. Use when asked I'm stuck on this and don't know why, I keep avoiding this task, how long should I struggle before asking, or unblock my stalled project. Produces the stuck diagnosis, the matched move, the ask-for-help script that preserves standing, and the stuck-log pattern read.
.claude/skills/mohitagw15856-unblock-protocol/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 55% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 72% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 99% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 72% | 0% |
"Stuck" is five different states wearing one feeling, and the wrong move for the type wastes days: struggling nobly on a don't-know-how (when a 10-minute ask solves it), asking for help on a can't-decide (when the decider is you), grinding on an avoiding (where the block is emotional, and effort slides off it). The protocol: diagnose the type first (one minute, five questions), apply the matched move, and honor the struggle-timebox — the pre-committed point where solo effort converts to an ask, because "I should be able to figure this out" is how competent people burn days protecting an image nobody's watching.
Ask for these if not provided:
The type, from the tells: (the symptom evidence) · the feeling noted as data]
The type-matched unblock, concretely: the timebox set / the decision forced / the chase+parallel / the shrink+name / the decomposition]
"Trying to X · tried A, B · A fails with Y · what am I missing?" → sent to (the right person)]
Type recorded · the pattern count · the structural note if this is entry #3 of a kind]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 21,404 | 23,446 | +10% | 1 | 1 | 0% | 2,154 | 3,346 | +55% | 0 | 0 | — |
case-02 | fail→pass | 17,535 | 23,371 | +33% | 1 | 1 | 0% | 2,396 | 4,118 | +72% | 0 | 0 | — |
case-03 | fail→pass | 23,396 | 17,621 | -25% | 1 | 1 | 0% | 2,383 | 3,206 | +35% | 0 | 0 | — |
case-04 | pass→fail | 30,690 | 30,412 | -1% | 1 | 1 | 0% | 5,375 | 6,360 | +18% | 0 | 0 | — |
case-05 | pass→fail | 23,007 | 28,218 | +23% | 1 | 1 | 0% | 2,258 | 4,997 | +121% | 0 | 0 | — |
case-06 | pass→pass | 24,675 | 42,145 | +71% | 1 | 1 | 0% | 3,100 | 6,532 | +111% | 0 | 0 | — |
case-07 | pass→fail | 20,816 | 22,353 | +7% | 1 | 1 | 0% | 2,139 | 4,446 | +108% | 0 | 0 | — |
case-08 | pass→pass | 17,541 | 14,316 | -18% | 1 | 1 | 0% | 1,813 | 2,812 | +55% | 0 | 0 | — |
case-09 | pass→pass | 20,845 | 17,279 | -17% | 1 | 1 | 0% | 1,912 | 2,886 | +51% | 0 | 0 | — |
case-10 | pass→pass | 18,575 | 17,709 | -5% | 1 | 1 | 0% | 2,079 | 2,967 | +43% | 0 | 0 | — |
case-17 | pass→pass | 16,739 | 15,153 | -9% | 1 | 1 | 0% | 1,626 | 2,532 | +56% | 0 | 0 | — |
case-11 | fail→fail | 13,780 | 17,131 | +24% | 1 | 1 | 0% | 2,092 | 3,219 | +54% | 0 | 0 | — |
case-12 | pass→pass | 18,543 | 15,483 | -17% | 1 | 1 | 0% | 2,042 | 2,952 | +45% | 0 | 0 | — |
case-13 | fail→pass | 16,942 | 12,623 | -25% | 1 | 1 | 0% | 1,642 | 3,260 | +99% | 0 | 0 | — |
case-14 | pass→pass | 8,744 | 15,320 | +75% | 1 | 1 | 0% | 1,449 | 2,699 | +86% | 0 | 0 | — |
case-15 | pass→pass | 14,524 | 17,577 | +21% | 1 | 1 | 0% | 1,363 | 3,192 | +134% | 0 | 0 | — |
case-16 | fail→pass | 17,287 | 15,314 | -11% | 1 | 1 | 0% | 1,679 | 2,885 | +72% | 0 | 0 | — |
case-18 | fail→pass | 14,744 | 13,628 | -8% | 1 | 1 | 0% | 1,415 | 2,263 | +60% | 0 | 0 | — |
case-19 | fail→fail | 14,787 | 13,969 | -6% | 1 | 1 | 0% | 1,502 | 2,742 | +83% | 0 | 0 | — |
case-20 | fail→fail | 23,730 | 17,678 | -26% | 1 | 1 | 0% | 2,611 | 3,715 | +42% | 0 | 0 | — |
case-21 | fail→pass | 26,580 | 22,953 | -14% | 1 | 1 | 0% | 2,643 | 3,685 | +39% | 0 | 0 | — |
case-22 | fail→fail | 18,269 | 9,512 | -48% | 1 | 1 | 0% | 1,978 | 2,713 | +37% | 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 +18 percentage points is the difference between those two pass rates over the 22 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.