Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →AI-powered autonomous extraction. The agent navigates sites and extracts structured data (takes 2-5 minutes).
.claude/skills/firecrawl-agent/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -49% | 0% |
AI-powered autonomous extraction. The agent navigates sites and extracts structured data (takes 2-5 minutes).
bash# Extract structured data firecrawl agent "extract all pricing tiers" --wait --json -o .firecrawl/pricing.json # With a JSON schema for structured output firecrawl agent "extract products" --schema '{"type":"object","properties":{"name":{"type":"string"},"price":{"type":"number"}}}' --wait --json -o .firecrawl/products.json # Focus on specific pages firecrawl agent "get feature list" --urls "<url>" --wait --json -o .firecrawl/features.json
Run firecrawl agent --help for the full option list.
Done when: the output file contains valid JSON answering the request — or a job ID was intentionally returned for later polling.
Omitting --wait returns a job ID. A UUID positional argument is auto-detected as a status check:
bash# Check once (equivalent to adding --status) firecrawl agent "<job-id>" # Wait on an existing job, polling every 10 seconds for up to 5 minutes firecrawl agent "<job-id>" --wait --poll-interval 10 --timeout 300 # Cancel an active job firecrawl agent "<job-id>" --cancel
--wait for inline results; omit it only when you want a job ID to poll later (see Job IDs).--schema for predictable, structured output — otherwise the agent returns freeform data.--max-credits to cap spending.scrape — it's faster and cheaper.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 16,348 | 8,864 | -46% | 1 | 1 | 0% | 2,333 | 857 | -63% | 0 | 0 | — |
case-15 | pass→pass | 12,980 | 7,922 | -39% | 1 | 1 | 0% | 1,059 | 876 | -17% | 0 | 0 | — |
case-14 | fail→pass | 11,695 | 9,203 | -21% | 1 | 1 | 0% | 1,883 | 917 | -51% | 0 | 0 | — |
case-02 | fail→pass | 19,344 | 17,552 | -9% | 1 | 1 | 0% | 885 | 1,075 | +21% | 0 | 0 | — |
case-03 | fail→pass | 18,782 | 7,070 | -62% | 1 | 1 | 0% | 602 | 925 | +54% | 0 | 0 | — |
case-04 | pass→pass | 9,153 | 8,204 | -10% | 1 | 1 | 0% | 734 | 1,110 | +51% | 0 | 0 | — |
case-05 | pass→pass | 17,154 | 10,019 | -42% | 1 | 1 | 0% | 2,075 | 1,446 | -30% | 0 | 0 | — |
case-06 | pass→pass | 22,105 | 10,589 | -52% | 1 | 1 | 0% | 2,279 | 1,494 | -34% | 0 | 0 | — |
case-07 | pass→pass | 17,002 | 16,180 | -5% | 1 | 1 | 0% | 1,961 | 2,153 | +10% | 0 | 0 | — |
case-08 | pass→pass | 14,286 | 7,778 | -46% | 1 | 1 | 0% | 1,499 | 1,099 | -27% | 0 | 0 | — |
case-09 | pass→pass | 18,431 | 7,095 | -62% | 1 | 1 | 0% | 2,315 | 928 | -60% | 0 | 0 | — |
case-10 | fail→pass | 13,546 | 7,047 | -48% | 1 | 1 | 0% | 1,587 | 816 | -49% | 0 | 0 | — |
case-11 | fail→pass | 14,191 | 7,387 | -48% | 1 | 1 | 0% | 1,422 | 922 | -35% | 0 | 0 | — |
case-12 | pass→pass | 10,469 | 9,381 | -10% | 1 | 1 | 0% | 1,047 | 1,353 | +29% | 0 | 0 | — |
case-13 | fail→pass | 18,194 | 10,537 | -42% | 1 | 1 | 0% | 2,266 | 1,646 | -27% | 0 | 0 | — |
case-16 | pass→pass | 28,939 | 8,145 | -72% | 1 | 1 | 0% | 3,712 | 930 | -75% | 0 | 0 | — |
case-17 | pass→pass | 18,013 | 8,767 | -51% | 1 | 1 | 0% | 2,558 | 1,134 | -56% | 0 | 0 | — |
case-18 | pass→pass | 18,855 | 10,429 | -45% | 1 | 1 | 0% | 2,095 | 1,369 | -35% | 0 | 0 | — |
case-19 | fail→pass | 14,875 | 10,080 | -32% | 1 | 1 | 0% | 1,447 | 1,260 | -13% | 0 | 0 | — |
case-20 | fail→pass | 28,599 | 2,467 | -91% | 1 | 1 | 0% | 1,959 | 829 | -58% | 0 | 0 | — |
case-21 | fail→pass | 18,357 | 7,724 | -58% | 1 | 1 | 0% | 2,651 | 900 | -66% | 0 | 0 | — |
case-22 | pass→pass | 5,058 | 1,967 | -61% | 1 | 1 | 0% | 938 | 734 | -22% | 0 | 0 | — |
case-23 | pass→pass | 18,082 | 2,854 | -84% | 1 | 1 | 0% | 1,828 | 949 | -48% | 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. 23 cases were attempted, and 22 counted toward the lift figure. The other 1 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 +43 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/23/2026 | +74% |
| gemini-3.6-flash | verified | 7/23/2026 | +77% |
Other measured skills in the registry, with their headline benchmark lift.