Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Fill long web forms and applications through a computer-use agent — from a fact sheet you approve, field by field, with a full transcript and nothing submitted without your word. Use when asked to fill this application for me, complete this government/vendor/insurance form, or do this registration. Produces the fact-to-field mapping, the filled form held at review, and a field-level transcript.
.claude/skills/mohitagw15856-form-filler-operator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 645% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 130% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 352% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 29% | 0% |
Forms are where an hour of a professional's day goes to die: the same facts, retyped into someone else's boxes. This skill runs the retyping through a computer-use agent — with the discipline the task actually needs: facts come from an approved sheet (never invented), every field write is logged, and the submit button belongs to the human. (Reading the form's fine print first? That's tos-decoder's job — chain them.)
Ask for these if not provided:
Fact sheet: n] facts confirmed · Fields: n] mapped, n] user-typed (sensitive), n] open questions | Field | Entered | Fact source | |---|---|---| Status: held at review — screenshot] — awaiting your submit.
tos-decoder) and put the checkbox in the user's handsFor computer-use agents with browser control. Without tools, the fact sheet + mapping is the deliverable (still worth an hour). Rules per SKILLSPEC.md §5.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 8,461 | 15,517 | +83% | 1 | 1 | 0% | 1,575 | 3,765 | +139% | 0 | 0 | — |
case-02 | fail→fail | 9,598 | 11,918 | +24% | 1 | 1 | 0% | 1,646 | 3,595 | +118% | 0 | 0 | — |
case-03 | fail→fail | 3,842 | 7,664 | +99% | 1 | 1 | 0% | 607 | 2,254 | +271% | 0 | 0 | — |
case-04 | pass→pass | 8,205 | 6,664 | -19% | 1 | 1 | 0% | 1,524 | 2,140 | +40% | 0 | 0 | — |
case-05 | pass→pass | 15,200 | 17,542 | +15% | 1 | 1 | 0% | 2,694 | 4,021 | +49% | 0 | 0 | — |
case-06 | fail→pass | 1,539 | 5,940 | +286% | 1 | 1 | 0% | 275 | 2,049 | +645% | 0 | 0 | — |
case-07 | fail→fail | 10,715 | 7,079 | -34% | 1 | 1 | 0% | 1,809 | 2,193 | +21% | 0 | 0 | — |
case-20 | fail→pass | 7,887 | 2,555 | -68% | 1 | 1 | 0% | 1,227 | 1,419 | +16% | 0 | 0 | — |
case-08 | fail→fail | 4,636 | 8,132 | +75% | 1 | 1 | 0% | 710 | 2,239 | +215% | 0 | 0 | — |
case-09 | pass→pass | 8,464 | 3,454 | -59% | 1 | 1 | 0% | 729 | 1,542 | +112% | 0 | 0 | — |
case-10 | pass→pass | 2,529 | 3,380 | +34% | 1 | 1 | 0% | 382 | 1,575 | +312% | 0 | 0 | — |
case-11 | fail→pass | 5,091 | 5,267 | +3% | 1 | 1 | 0% | 777 | 1,788 | +130% | 0 | 0 | — |
case-12 | fail→fail | 6,688 | 4,702 | -30% | 1 | 1 | 0% | 1,048 | 1,810 | +73% | 0 | 0 | — |
case-13 | fail→fail | 5,480 | 5,746 | +5% | 1 | 1 | 0% | 870 | 1,878 | +116% | 0 | 0 | — |
case-14 | pass→pass | 5,717 | 6,896 | +21% | 1 | 1 | 0% | 850 | 2,165 | +155% | 0 | 0 | — |
case-15 | pass→pass | 3,798 | 4,700 | +24% | 1 | 1 | 0% | 580 | 1,730 | +198% | 0 | 0 | — |
case-21 | pass→pass | 9,845 | 7,146 | -27% | 1 | 1 | 0% | 1,572 | 2,208 | +40% | 0 | 0 | — |
case-16 | pass→pass | 10,094 | 6,135 | -39% | 1 | 1 | 0% | 1,558 | 1,985 | +27% | 0 | 0 | — |
case-17 | fail→pass | 3,018 | 6,380 | +111% | 1 | 1 | 0% | 448 | 2,024 | +352% | 0 | 0 | — |
case-18 | pass→pass | 8,602 | 4,271 | -50% | 1 | 1 | 0% | 1,343 | 1,693 | +26% | 0 | 0 | — |
case-19 | fail→pass | 11,192 | 7,874 | -30% | 1 | 1 | 0% | 1,964 | 2,531 | +29% | 0 | 0 | — |
case-22 | fail→pass | 5,753 | 2,442 | -58% | 1 | 1 | 0% | 927 | 1,403 | +51% | 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 +27 percentage points is the difference between those two pass rates over the 22 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.