Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Execute Firecrawl production deployment checklist and rollback procedures. Use when deploying Firecrawl integrations to production, preparing for launch, or implementing go-live procedures. Trigger with phrases like "firecrawl production", "deploy firecrawl", "firecrawl go-live", "firecrawl launch checklist".
.claude/skills/jeremylongshore-firecrawl-prod-checklist/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 78% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 71% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -9% | 0% |
Produce evidence for a go/no-go decision. A checked box without an artifact, test, owner, or exact release reference is not evidence.
The release must use current v2 endpoints or an explicitly isolated legacy path, the official SDK surface, complete async pagination, documented retryability, signed webhooks, explicit crawl/batch limits, current billing controls, and approved cache/retention behavior. Self-hosted releases require their own production controls beyond Compose evaluation.
For authenticated Cloud operations, inject FIRECRAWL_API_KEY from an approved secret manager. REST requests use Authorization: Bearer with the key. Never print, commit, transmit, or place a key in a URL. Keyless access is suitable only where the current documentation explicitly allows it and the workload accepts its limits; production workflows should make identity and team ownership explicit.
Use Read, Glob, and Grep to inspect code, configuration, tests, and evidence. Use Write/Edit only for approved implementation or documentation changes. Do not call Firecrawl, rotate keys, change account settings, scrape a target, or deploy merely because this skill was invoked.
Require named security/data/product/operations approval for their controls and release-owner approval for go-live, plan/spend changes, exceptions, and rollback waivers.
Return a release scorecard with evidence links, exact versions, gate verdicts, canary and rollback results, exceptions, owners, expiry, and final go/no-go decision.
Read official Firecrawl evidence before relying on an endpoint, SDK method, plan limit, price, retention option, or self-hosted release.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 23,240 | 31,401 | +35% | 1 | 1 | 0% | 3,067 | 6,353 | +107% | 0 | 0 | — |
case-02 | fail→pass | 31,982 | 20,483 | -36% | 1 | 1 | 0% | 5,507 | 4,841 | -12% | 0 | 0 | — |
case-03 | fail→pass | 19,305 | 29,519 | +53% | 1 | 1 | 0% | 3,368 | 5,999 | +78% | 0 | 0 | — |
case-04 | fail→fail | 11,436 | 10,938 | -4% | 1 | 1 | 0% | 2,153 | 2,502 | +16% | 0 | 0 | — |
case-05 | fail→pass | 10,642 | 8,308 | -22% | 1 | 1 | 0% | 1,189 | 2,039 | +71% | 0 | 0 | — |
case-06 | pass→pass | 9,959 | 2,313 | -77% | 1 | 1 | 0% | 1,560 | 1,730 | +11% | 0 | 0 | — |
case-07 | fail→fail | 24,505 | 14,384 | -41% | 1 | 1 | 0% | 3,480 | 3,153 | -9% | 0 | 0 | — |
case-08 | pass→pass | 16,110 | 10,354 | -36% | 1 | 1 | 0% | 3,028 | 2,350 | -22% | 0 | 0 | — |
case-09 | pass→pass | 7,913 | 7,207 | -9% | 1 | 1 | 0% | 1,378 | 1,704 | +24% | 0 | 0 | — |
case-10 | pass→pass | 19,894 | 18,196 | -9% | 1 | 1 | 0% | 2,721 | 4,258 | +56% | 0 | 0 | — |
case-11 | fail→fail | 14,751 | 15,716 | +7% | 1 | 1 | 0% | 2,502 | 4,153 | +66% | 0 | 0 | — |
case-12 | pass→pass | 19,009 | 2,963 | -84% | 1 | 1 | 0% | 2,273 | 1,912 | -16% | 0 | 0 | — |
case-13 | fail→pass | 17,370 | 11,533 | -34% | 1 | 1 | 0% | 2,949 | 3,504 | +19% | 0 | 0 | — |
case-14 | pass→pass | 18,891 | 11,159 | -41% | 1 | 1 | 0% | 2,212 | 2,477 | +12% | 0 | 0 | — |
case-15 | pass→pass | 21,437 | 4,754 | -78% | 1 | 1 | 0% | 2,889 | 2,107 | -27% | 0 | 0 | — |
case-16 | fail→pass | 19,794 | 11,102 | -44% | 1 | 1 | 0% | 2,750 | 2,510 | -9% | 0 | 0 | — |
case-17 | pass→pass | 17,830 | 8,074 | -55% | 1 | 1 | 0% | 2,336 | 2,751 | +18% | 0 | 0 | — |
case-18 | fail→pass | 11,976 | 6,397 | -47% | 1 | 1 | 0% | 1,211 | 1,567 | +29% | 0 | 0 | — |
case-19 | pass→pass | 20,321 | 19,846 | -2% | 1 | 1 | 0% | 2,640 | 4,144 | +57% | 0 | 0 | — |
case-20 | pass→pass | 22,870 | 19,510 | -15% | 1 | 1 | 0% | 4,395 | 5,213 | +19% | 0 | 0 | — |
case-21 | pass→pass | 18,529 | 16,139 | -13% | 1 | 1 | 0% | 2,795 | 4,745 | +70% | 0 | 0 | — |
case-22 | pass→pass | 16,652 | 15,773 | -5% | 1 | 1 | 0% | 3,955 | 5,136 | +30% | 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.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
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.