Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Simulate the adjuster's settlement call after your accident or loss — the recorded-statement asks, the quick-settlement anchor, the friendly minimization — run against your actual claim, with a debrief on every answer that shrank it. Use when asked the adjuster wants a recorded statement, practice the settlement call, is this settlement offer low, or what will the insurance company try. Produces the call transcript with the adjuster's file notes, the offer trajectory, and a debrief separating fa
.claude/skills/mohitagw15856-the-insurance-adjuster/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 110% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 121% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 136% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 71% | 0% |
The adjuster on the phone is professionally friendly and professionally adverse — a claims handler whose file notes score everything you say against payout-reducing categories: pre-existing, minimizing language, gaps in treatment, comparative fault. This skill plays that call against your actual claim — the recorded-statement request, the "how are you feeling today?" that becomes "claimant reports feeling fine," the fast low anchor — and debriefs which answers helped the file and which quietly shrank it. It trains the call; it is not legal advice, and one of its jobs is naming when the situation has outgrown a phone call and needs a lawyer.
Ask for these if not provided:
> Simulation — a plausible adversarial reading, not a prediction.
The call. File note: after each answer — what entered the record and which category it fed]
Anchor → movements → what the notes say was actually available · the release-attached warning where it appears]
| Your answer | File effect | The safe version | |---|---|---| Plus the bright lines: no recorded statement for the other side without consideration · no medical self-assessment — route to records · no fault discussion — facts only · no settlement talk before treatment completes · everything material in writing]
> Training for a phone call, not legal advice — policy duties vary, releases are permanent, and the trigger list above is where a lawyer stops being optional. Consultations with injury lawyers are typically free; use one past any trigger.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 55,535 | 52,008 | -6% | 1 | 1 | 0% | 4,164 | 8,750 | +110% | 0 | 0 | — |
case-02 | fail→pass | 39,510 | 41,090 | +4% | 1 | 1 | 0% | 4,527 | 6,182 | +37% | 0 | 0 | — |
case-03 | pass→pass | 38,891 | 33,486 | -14% | 1 | 1 | 0% | 4,762 | 5,809 | +22% | 0 | 0 | — |
case-04 | pass→pass | 24,439 | 34,625 | +42% | 1 | 1 | 0% | 2,861 | 5,031 | +76% | 0 | 0 | — |
case-05 | pass→pass | 28,147 | 31,547 | +12% | 1 | 1 | 0% | 3,089 | 5,050 | +63% | 0 | 0 | — |
case-12 | pass→pass | 13,082 | 34,976 | +167% | 1 | 1 | 0% | 1,850 | 6,413 | +247% | 0 | 0 | — |
case-06 | pass→pass | 34,615 | 32,455 | -6% | 1 | 1 | 0% | 3,302 | 5,174 | +57% | 0 | 0 | — |
case-07 | fail→pass | 45,815 | 43,881 | -4% | 1 | 1 | 0% | 3,165 | 6,994 | +121% | 0 | 0 | — |
case-08 | pass→pass | 17,904 | 30,950 | +73% | 1 | 1 | 0% | 2,679 | 5,280 | +97% | 0 | 0 | — |
case-09 | pass→pass | 23,702 | 32,204 | +36% | 1 | 1 | 0% | 3,021 | 4,722 | +56% | 0 | 0 | — |
case-10 | fail→pass | 15,355 | 36,936 | +141% | 1 | 1 | 0% | 2,280 | 5,382 | +136% | 0 | 0 | — |
case-11 | fail→fail | 43,070 | 32,938 | -24% | 1 | 1 | 0% | 3,326 | 5,050 | +52% | 0 | 0 | — |
case-13 | pass→pass | 9,610 | 27,363 | +185% | 1 | 1 | 0% | 1,517 | 4,597 | +203% | 0 | 0 | — |
case-14 | fail→pass | 26,046 | 31,357 | +20% | 1 | 1 | 0% | 2,989 | 5,099 | +71% | 0 | 0 | — |
case-15 | pass→pass | 19,119 | 37,434 | +96% | 1 | 1 | 0% | 2,057 | 6,843 | +233% | 0 | 0 | — |
case-16 | fail→pass | 8,180 | 14,915 | +82% | 1 | 1 | 0% | 960 | 2,554 | +166% | 0 | 0 | — |
case-17 | pass→pass | 22,252 | 44,344 | +99% | 1 | 1 | 0% | 2,464 | 6,860 | +178% | 0 | 0 | — |
case-18 | pass→pass | 20,069 | 28,367 | +41% | 1 | 1 | 0% | 2,314 | 4,649 | +101% | 0 | 0 | — |
case-19 | pass→pass | 26,297 | 41,499 | +58% | 1 | 1 | 0% | 3,081 | 6,011 | +95% | 0 | 0 | — |
case-20 | pass→pass | 21,270 | 38,591 | +81% | 1 | 1 | 0% | 2,844 | 4,912 | +73% | 0 | 0 | — |
case-21 | pass→fail | 25,783 | 41,300 | +60% | 1 | 1 | 0% | 3,051 | 7,355 | +141% | 0 | 0 | — |
case-22 | pass→pass | 26,955 | 24,932 | -8% | 1 | 1 | 0% | 3,270 | 4,566 | +40% | 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 +23 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.