Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a defensible salary range for a role — what it actually pays given the market, location, level, and your value — so you can ask, counter, or set pay with a real number. Use when asked what should I be paid, is my salary fair, research market pay for [role], or how much to ask for. Produces a structured way to research the range from multiple sources, the factors that move your number (level, location, industry, skills, company size), where you likely sit in the band, and how to frame the n
.claude/skills/mohitagw15856-salary-benchmarking/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 297% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 33% | 0% |
"Am I underpaid?" and "what should I ask for?" both need the same thing: a defensible number, not a guess or a single glassdoor figure. This builds that range by triangulating sources and adjusting for the factors that actually move pay — level, location, industry, skills, company size — then places you within the band and helps you frame it. (Making the ask itself is a separate negotiation skill.)
Ask for these if not provided:
Research from: surveys · aggregators · live postings · peers · recruiters]. Adjust for: level x] · location/COL y] · industry · scarce skills · company size/stage. Your position in the band: where + why — experience/skills/results]. Defensible range: low ] · target ] · high ] — justified by factors]. How to frame it: number + evidence, tied to value].
> Pay data is noisy and can be stale — triangulate multiple sources and adjust for your specifics. (Running the negotiation itself is a separate step.)
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 36,879 | 26,842 | -27% | 1 | 1 | 0% | 5,376 | 3,923 | -27% | 0 | 0 | — |
case-02 | fail→fail | 35,538 | 20,632 | -42% | 1 | 1 | 0% | 5,389 | 3,599 | -33% | 0 | 0 | — |
case-03 | fail→pass | 35,273 | 19,743 | -44% | 1 | 1 | 0% | 4,913 | 4,531 | -8% | 0 | 0 | — |
case-04 | fail→fail | 24,616 | 19,329 | -21% | 1 | 1 | 0% | 3,162 | 3,502 | +11% | 0 | 0 | — |
case-05 | fail→fail | 23,431 | 18,483 | -21% | 1 | 1 | 0% | 2,426 | 3,070 | +27% | 0 | 0 | — |
case-06 | fail→pass | 20,489 | 12,242 | -40% | 1 | 1 | 0% | 2,749 | 2,107 | -23% | 0 | 0 | — |
case-07 | pass→pass | 23,109 | 21,999 | -5% | 1 | 1 | 0% | 2,448 | 3,285 | +34% | 0 | 0 | — |
case-08 | pass→pass | 13,107 | 21,543 | +64% | 1 | 1 | 0% | 2,136 | 3,297 | +54% | 0 | 0 | — |
case-09 | pass→pass | 20,652 | 18,210 | -12% | 1 | 1 | 0% | 2,242 | 3,098 | +38% | 0 | 0 | — |
case-10 | fail→fail | 15,814 | 16,017 | +1% | 1 | 1 | 0% | 2,446 | 2,736 | +12% | 0 | 0 | — |
case-11 | pass→pass | 14,029 | 19,643 | +40% | 1 | 1 | 0% | 2,290 | 3,368 | +47% | 0 | 0 | — |
case-12 | pass→pass | 18,438 | 20,997 | +14% | 1 | 1 | 0% | 2,307 | 3,567 | +55% | 0 | 0 | — |
case-13 | fail→pass | 9,724 | 12,116 | +25% | 1 | 1 | 0% | 756 | 3,003 | +297% | 0 | 0 | — |
case-14 | pass→pass | 18,925 | 13,936 | -26% | 1 | 1 | 0% | 2,192 | 2,252 | +3% | 0 | 0 | — |
case-15 | pass→pass | 18,353 | 17,631 | -4% | 1 | 1 | 0% | 1,960 | 2,705 | +38% | 0 | 0 | — |
case-16 | pass→pass | 24,723 | 20,020 | -19% | 1 | 1 | 0% | 2,500 | 3,381 | +35% | 0 | 0 | — |
case-17 | pass→pass | 16,679 | 16,040 | -4% | 1 | 1 | 0% | 2,308 | 3,455 | +50% | 0 | 0 | — |
case-18 | pass→pass | 13,056 | 3,938 | -70% | 1 | 1 | 0% | 1,311 | 1,579 | +20% | 0 | 0 | — |
case-19 | pass→pass | 24,517 | 16,135 | -34% | 1 | 1 | 0% | 2,549 | 3,436 | +35% | 0 | 0 | — |
case-20 | fail→pass | 21,975 | 15,782 | -28% | 1 | 1 | 0% | 2,544 | 3,393 | +33% | 0 | 0 | — |
case-21 | fail→pass | 18,920 | 21,023 | +11% | 1 | 1 | 0% | 2,618 | 3,677 | +40% | 0 | 0 | — |
case-22 | pass→pass | 13,808 | 13,049 | -5% | 1 | 1 | 0% | 2,175 | 3,020 | +39% | 0 | 0 | — |
case-23 | fail→pass | 17,481 | 25,243 | +44% | 1 | 1 | 0% | 1,980 | 3,553 | +79% | 0 | 0 | — |
case-24 | pass→pass | 21,180 | 21,955 | +4% | 1 | 1 | 0% | 3,563 | 3,624 | +2% | 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 +29 percentage points is the difference between those two pass rates over the 24 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.