Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Turn an anxious spiral into a concrete list — separate the specific worries from the vague dread, sort what you can act on from what you can't, and get one action. Use when asked help me with my anxiety spiral, I can't stop worrying about, my mind won't stop racing, or break down what I'm anxious about. Produces the swirling worry pulled apart into named, specific concerns, each sorted into can-act-on vs can't-control vs not-actually-likely, a single action for the actionable ones, and a way to
.claude/skills/mohitagw15856-the-worry-decompiler/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 106% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 59% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 34% | 0% |
A worry spiral is one giant, formless dread looping on itself — impossible to act on because it's not a thing, it's a fog. This decompiles it: pulls the swirl apart into specific, named worries, sorts each into what you can actually do something about vs. what you can't control vs. what isn't even likely, hands you one action for the actionable, and helps you set down the rest. A list is manageable in a way a spiral never is. It's a structuring tool, not therapy.
Ask for these if not provided:
Sorted
👉 The one action: concrete next step for the top actionable worry]. If it's physical right now: a simple grounding technique].
> A way to make anxiety manageable, not therapy. If the spirals are frequent or overwhelming, please talk to someone you trust or a professional.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 13,067 | 12,312 | -6% | 1 | 1 | 0% | 1,021 | 2,100 | +106% | 0 | 0 | — |
case-02 | pass→pass | 14,044 | 7,664 | -45% | 1 | 1 | 0% | 1,132 | 1,959 | +73% | 0 | 0 | — |
case-03 | fail→fail | 11,902 | 14,769 | +24% | 1 | 1 | 0% | 947 | 2,412 | +155% | 0 | 0 | — |
case-04 | pass→pass | 14,918 | 16,709 | +12% | 1 | 1 | 0% | 2,513 | 3,300 | +31% | 0 | 0 | — |
case-05 | pass→pass | 19,219 | 19,695 | +2% | 1 | 1 | 0% | 2,477 | 3,025 | +22% | 0 | 0 | — |
case-06 | pass→pass | 17,506 | 17,366 | -1% | 1 | 1 | 0% | 2,467 | 2,965 | +20% | 0 | 0 | — |
case-07 | pass→fail | 16,372 | 18,159 | +11% | 1 | 1 | 0% | 1,810 | 3,383 | +87% | 0 | 0 | — |
case-08 | pass→pass | 35,446 | 15,391 | -57% | 1 | 1 | 0% | 1,942 | 2,537 | +31% | 0 | 0 | — |
case-09 | fail→fail | 17,328 | 17,029 | -2% | 1 | 1 | 0% | 2,105 | 2,974 | +41% | 0 | 0 | — |
case-10 | pass→pass | 20,964 | 20,194 | -4% | 1 | 1 | 0% | 2,178 | 3,116 | +43% | 0 | 0 | — |
case-11 | fail→pass | 19,931 | 18,696 | -6% | 1 | 1 | 0% | 2,331 | 2,792 | +20% | 0 | 0 | — |
case-12 | fail→fail | 18,448 | 14,505 | -21% | 1 | 1 | 0% | 1,666 | 2,306 | +38% | 0 | 0 | — |
case-13 | fail→pass | 25,973 | 17,269 | -34% | 1 | 1 | 0% | 3,031 | 3,186 | +5% | 0 | 0 | — |
case-14 | fail→pass | 20,163 | 14,470 | -28% | 1 | 1 | 0% | 1,899 | 3,018 | +59% | 0 | 0 | — |
case-15 | fail→pass | 20,861 | 17,366 | -17% | 1 | 1 | 0% | 2,195 | 2,950 | +34% | 0 | 0 | — |
case-16 | pass→pass | 23,190 | 16,335 | -30% | 1 | 1 | 0% | 2,176 | 2,758 | +27% | 0 | 0 | — |
case-17 | pass→pass | 26,241 | 20,283 | -23% | 1 | 1 | 0% | 2,585 | 2,903 | +12% | 0 | 0 | — |
case-18 | fail→fail | 15,255 | 12,948 | -15% | 1 | 1 | 0% | 1,480 | 2,772 | +87% | 0 | 0 | — |
case-19 | fail→pass | 23,336 | 15,035 | -36% | 1 | 1 | 0% | 2,867 | 2,795 | -3% | 0 | 0 | — |
case-20 | pass→pass | 23,658 | 15,186 | -36% | 1 | 1 | 0% | 2,424 | 2,629 | +8% | 0 | 0 | — |
case-21 | fail→pass | 20,413 | 15,276 | -25% | 1 | 1 | 0% | 1,887 | 2,571 | +36% | 0 | 0 | — |
case-22 | fail→pass | 10,910 | 15,698 | +44% | 1 | 1 | 0% | 1,634 | 2,709 | +66% | 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 +32 percentage points is the difference between those two pass rates over the 22 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.