Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Scrape any webpage as clean markdown via Bright Data Web Unlocker API. Bypasses bot detection and CAPTCHA. Requires BRIGHTDATA_API_KEY and BRIGHTDATA_UNLOCKER_ZONE environment variables.
.claude/skills/davila7-scrape/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -64% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -48% | 0% |
Scrape any webpage and get clean markdown content using Bright Data's Web Unlocker API. Automatically bypasses bot detection and CAPTCHA.
1. Get your API Key: Get a key from Bright Data Dashboard.
2. Create a Web Unlocker zone: Create a zone at brightdata.com/cp by clicking "Add" (top-right), selecting "Unlocker zone".
3. Set environment variables:
bashexport BRIGHTDATA_API_KEY="your-api-key" export BRIGHTDATA_UNLOCKER_ZONE="your-zone-name"
bashbash scripts/scrape.sh "url"
Parameters:
url (required): The webpage URL to scrapeExamples:
bash# Scrape a news article bash scripts/scrape.sh "https://example.com/article" # Scrape a product page bash scripts/scrape.sh "https://shop.example.com/product/123"
Returns clean markdown content extracted from the webpage:
markdown# Page Title Main content of the page converted to markdown format... ## Section Heading More content...
curl - For API requests| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 4,637 | 1,893 | -59% | 1 | 1 | 0% | 839 | 638 | -24% | 0 | 0 | — |
case-02 | fail→fail | 4,054 | 5,360 | +32% | 1 | 1 | 0% | 643 | 572 | -11% | 0 | 0 | — |
case-03 | fail→pass | 8,815 | 4,959 | -44% | 1 | 1 | 0% | 1,649 | 596 | -64% | 0 | 0 | — |
case-04 | fail→fail | 7,940 | 2,434 | -69% | 1 | 1 | 0% | 1,306 | 698 | -47% | 0 | 0 | — |
case-05 | fail→pass | 7,362 | 3,404 | -54% | 1 | 1 | 0% | 1,139 | 701 | -38% | 0 | 0 | — |
case-06 | pass→pass | 5,357 | 2,123 | -60% | 1 | 1 | 0% | 766 | 582 | -24% | 0 | 0 | — |
case-07 | fail→fail | 15,128 | 12,013 | -21% | 1 | 1 | 0% | 1,804 | 1,744 | -3% | 0 | 0 | — |
case-08 | fail→pass | 10,951 | 4,622 | -58% | 1 | 1 | 0% | 1,777 | 1,112 | -37% | 0 | 0 | — |
case-09 | fail→pass | 8,425 | 2,635 | -69% | 1 | 1 | 0% | 1,401 | 733 | -48% | 0 | 0 | — |
case-10 | pass→pass | 6,332 | 1,577 | -75% | 1 | 1 | 0% | 1,040 | 538 | -48% | 0 | 0 | — |
case-11 | fail→pass | 7,692 | 2,138 | -72% | 1 | 1 | 0% | 1,323 | 688 | -48% | 0 | 0 | — |
case-12 | fail→pass | 5,553 | 2,833 | -49% | 1 | 1 | 0% | 1,037 | 673 | -35% | 0 | 0 | — |
case-13 | fail→pass | 8,413 | 2,602 | -69% | 1 | 1 | 0% | 1,360 | 731 | -46% | 0 | 0 | — |
case-14 | pass→pass | 7,905 | 3,373 | -57% | 1 | 1 | 0% | 1,198 | 919 | -23% | 0 | 0 | — |
case-15 | fail→pass | 15,288 | 11,841 | -23% | 1 | 1 | 0% | 1,794 | 1,335 | -26% | 0 | 0 | — |
case-16 | fail→pass | 10,813 | 1,332 | -88% | 1 | 1 | 0% | 1,578 | 511 | -68% | 0 | 0 | — |
case-17 | pass→fail | 15,256 | 5,888 | -61% | 1 | 1 | 0% | 2,418 | 591 | -76% | 0 | 0 | — |
case-18 | fail→fail | 4,499 | 8,761 | +95% | 1 | 1 | 0% | 856 | 1,109 | +30% | 0 | 0 | — |
case-19 | fail→pass | 5,308 | 2,585 | -51% | 1 | 1 | 0% | 838 | 562 | -33% | 0 | 0 | — |
case-20 | pass→pass | 8,131 | 6,286 | -23% | 1 | 1 | 0% | 1,501 | 1,561 | +4% | 0 | 0 | — |
case-21 | pass→pass | 9,327 | 4,691 | -50% | 1 | 1 | 0% | 1,699 | 1,226 | -28% | 0 | 0 | — |
case-22 | pass→pass | 5,991 | 3,752 | -37% | 1 | 1 | 0% | 1,026 | 862 | -16% | 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, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +41 percentage points is the difference between those two pass rates over the 19 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.