Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Bypasses anti-bot protections using Oxylabs Web Unblocker, an AI-powered proxy that handles fingerprinting, JavaScript rendering, and retries automatically. Use when the user needs to scrape protected websites, bypass CAPTCHAs, access blocked content, or when regular proxies fail due to anti-bot measures.
.claude/skills/oxylabs-web-unblocker/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 64% | 0% |
AI-powered proxy solution that automatically manages fingerprinting, headers, retries, and JavaScript rendering.
https://unblock.oxylabs.io:60000HTTP Basic Auth via proxy credentials:
bashcurl -k -x "https://unblock.oxylabs.io:60000" \ -U "$OXYLABS_USERNAME:$OXYLABS_PASSWORD" \ "https://example.com"
Basic request:
bashcurl -k -x "https://unblock.oxylabs.io:60000" \ -U "$OXYLABS_USERNAME:$OXYLABS_PASSWORD" \ "https://ip.oxylabs.io/headers"
With JavaScript rendering:
bashcurl -k -x "https://unblock.oxylabs.io:60000" \ -U "$OXYLABS_USERNAME:$OXYLABS_PASSWORD" \ -H "x-oxylabs-render: html" \ "https://example.com/spa-page"
| Header | Description | |--------|-------------| | x-oxylabs-render | html for rendered HTML, png for raw PNG bytes; empty value disables automatic forced rendering | | X-Oxylabs-Session-Id | Reuse same IP across requests (any random string) | | X-Oxylabs-Geo-Location | Target country, city/state, ZIP/postcode, or coordinates | | x-oxylabs-force-headers: 1 | Enable custom header passthrough | | x-oxylabs-force-cookies: 1 | Enable custom cookie passthrough | | X-Oxylabs-Successful-Status-Codes | Define custom success codes to prevent retries | | x-oxylabs-browser-instructions | JSON-escaped browser actions; requires x-oxylabs-render: html |
Reuse the same IP across multiple requests:
bashcurl -k -x "https://unblock.oxylabs.io:60000" \ -U "$OXYLABS_USERNAME:$OXYLABS_PASSWORD" \ -H "X-Oxylabs-Session-Id: my-session-123" \ "https://example.com/page1"
bashcurl -k -x "https://unblock.oxylabs.io:60000" \ -U "$OXYLABS_USERNAME:$OXYLABS_PASSWORD" \ -H "X-Oxylabs-Geo-Location: Germany" \ "https://example.com"
Use values such as Germany, 90210, California,United States, New York,New York,United States, or lat: 40.7128, lng: -74.0060, rad: 50.
Use normal HTTP methods and request bodies through the proxy; Web Unblocker supports both GET and POST.
| Scenario | Use | |----------|-----| | Sites with anti-bot protection | Web Unblocker | | CAPTCHAs, fingerprint detection | Web Unblocker | | JavaScript-heavy SPAs | Web Unblocker with x-oxylabs-render: html | | Simple requests, no protection | Regular Proxies | | High volume, price sensitive | Regular Proxies |
-k flag (or disable SSL verification) - the proxy uses its own certificatesx-oxylabs-render: html if experiencing empty content or low success rates; set client timeouts near 180 seconds for rendered requestsX-Oxylabs-Final-Url in response headers when redirects matterx-oxylabs-render: htmlFor code examples in Python, Node.js, PHP, Go, Java, and C#, see examples.md.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 22,314 | 25,679 | +15% | 1 | 1 | 0% | 1,305 | 1,706 | +31% | 0 | 0 | — |
case-02 | fail→pass | 12,147 | 5,079 | -58% | 1 | 1 | 0% | 1,898 | 1,885 | -1% | 0 | 0 | — |
case-03 | fail→pass | 12,286 | 12,852 | +5% | 1 | 1 | 0% | 1,953 | 1,920 | -2% | 0 | 0 | — |
case-04 | fail→pass | 18,853 | 4,785 | -75% | 1 | 1 | 0% | 2,887 | 1,736 | -40% | 0 | 0 | — |
case-05 | pass→pass | 14,050 | 4,269 | -70% | 1 | 1 | 0% | 2,413 | 1,489 | -38% | 0 | 0 | — |
case-06 | pass→pass | 8,141 | 3,487 | -57% | 1 | 1 | 0% | 1,307 | 1,448 | +11% | 0 | 0 | — |
case-07 | fail→pass | 19,230 | 3,507 | -82% | 1 | 1 | 0% | 950 | 1,559 | +64% | 0 | 0 | — |
case-08 | pass→pass | 6,754 | 5,957 | -12% | 1 | 1 | 0% | 870 | 1,239 | +42% | 0 | 0 | — |
case-09 | fail→pass | 13,235 | 3,227 | -76% | 1 | 1 | 0% | 2,025 | 1,345 | -34% | 0 | 0 | — |
case-10 | pass→pass | 17,513 | 13,062 | -25% | 1 | 1 | 0% | 2,404 | 2,836 | +18% | 0 | 0 | — |
case-11 | pass→pass | 18,949 | 5,865 | -69% | 1 | 1 | 0% | 2,107 | 1,870 | -11% | 0 | 0 | — |
case-12 | pass→pass | 8,384 | 2,827 | -66% | 1 | 1 | 0% | 688 | 1,327 | +93% | 0 | 0 | — |
case-13 | pass→pass | 9,053 | 10,080 | +11% | 1 | 1 | 0% | 1,594 | 1,665 | +4% | 0 | 0 | — |
case-14 | pass→pass | 5,884 | 2,699 | -54% | 1 | 1 | 0% | 844 | 1,270 | +50% | 0 | 0 | — |
case-15 | fail→pass | 10,301 | 3,838 | -63% | 1 | 1 | 0% | 1,825 | 1,413 | -23% | 0 | 0 | — |
case-16 | fail→pass | 5,889 | 3,519 | -40% | 1 | 1 | 0% | 889 | 1,386 | +56% | 0 | 0 | — |
case-17 | pass→pass | 25,177 | 4,698 | -81% | 1 | 1 | 0% | 2,234 | 1,623 | -27% | 0 | 0 | — |
case-18 | pass→pass | 7,689 | 8,590 | +12% | 1 | 1 | 0% | 1,133 | 1,387 | +22% | 0 | 0 | — |
case-19 | pass→pass | 4,373 | 2,365 | -46% | 1 | 1 | 0% | 643 | 1,227 | +91% | 0 | 0 | — |
case-20 | pass→pass | 9,611 | 4,175 | -57% | 1 | 1 | 0% | 1,770 | 1,510 | -15% | 0 | 0 | — |
case-21 | pass→pass | 11,820 | 3,657 | -69% | 1 | 1 | 0% | 1,562 | 1,277 | -18% | 0 | 0 | — |
case-22 | pass→pass | 8,952 | 9,757 | +9% | 1 | 1 | 0% | 1,490 | 2,299 | +54% | 0 | 0 | — |
case-23 | pass→pass | 8,864 | 7,282 | -18% | 1 | 1 | 0% | 1,515 | 2,175 | +44% | 0 | 0 | — |
case-24 | pass→pass | 7,143 | 6,360 | -11% | 1 | 1 | 0% | 1,078 | 1,581 | +47% | 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, and 22 counted toward the lift figure. The other 2 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 +33 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.