Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Execute Bright Data production deployment checklist and rollback procedures. Use when deploying Bright Data integrations to production, preparing for launch, or implementing go-live procedures. Trigger with phrases like "brightdata production", "deploy brightdata", "brightdata go-live", "brightdata launch checklist".
.claude/skills/jeremylongshore-brightdata-prod-checklist/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 49% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -20% | 0% |
Make production promotion a fail-closed evidence decision. A green connectivity test is insufficient without target authorization, schema and data controls, secret ownership, backpressure, observability, and a tested rollback.
Read approvals and Grep runtime configuration for unlisted targets, fields, products, zones, or delivery destinations. Verify public-data scope and recipient and retention decisions.
Confirm secret-manager binding, environment isolation, egress allowlists, redirect denial, byte, record, and job ceilings, adaptive backpressure, redacted telemetry, and idempotent processing.
Write or Edit tests for revoked credentials, policy 403, 429, provider 5xx, snapshot failure or expiry, malformed schema, oversized delivery, duplicate delivery, and downstream outage.
Run a bounded approved canary through the deployment workflow. Compare authorization, success, error, latency, cost, and data-quality thresholds; promote only with a rollback receipt.
Use Read and Grep for release-evidence inspection. Use Write and Edit for missing tests, manifests, runbooks, and the decision record. This skill does not change production credentials, zones, traffic, or delivery destinations.
Release one worker with a low job and byte ceiling and an approved test target. Promote only after duplicate delivery, policy denial, secret redaction, schema quarantine, and rollback drills all produce their expected receipts.
| Failure | Meaning | Response | |---------|---------|----------| | A required owner has not approved | Release authority is incomplete | Keep the candidate staged | | Rollback depends on an untested old zone | Recovery is speculative | Run the rollback drill before promotion | | Canary produces unexpected fields | Data contract drifted | Quarantine results and fail the gate |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 19,854 | 12,183 | -39% | 1 | 1 | 0% | 3,310 | 3,305 | -0% | 0 | 0 | — |
case-02 | fail→fail | 21,804 | 14,765 | -32% | 1 | 1 | 0% | 3,565 | 3,757 | +5% | 0 | 0 | — |
case-03 | fail→fail | 20,077 | 25,939 | +29% | 1 | 1 | 0% | 3,110 | 3,728 | +20% | 0 | 0 | — |
case-04 | fail→pass | 24,484 | 1,835 | -93% | 1 | 1 | 0% | 2,221 | 1,241 | -44% | 0 | 0 | — |
case-05 | fail→pass | 7,352 | 5,897 | -20% | 1 | 1 | 0% | 1,411 | 2,102 | +49% | 0 | 0 | — |
case-06 | pass→pass | 10,933 | 4,932 | -55% | 1 | 1 | 0% | 1,689 | 1,862 | +10% | 0 | 0 | — |
case-07 | pass→pass | 6,737 | 2,530 | -62% | 1 | 1 | 0% | 1,209 | 1,367 | +13% | 0 | 0 | — |
case-08 | fail→pass | 14,209 | 3,332 | -77% | 1 | 1 | 0% | 2,467 | 1,537 | -38% | 0 | 0 | — |
case-09 | fail→pass | 13,008 | 5,592 | -57% | 1 | 1 | 0% | 2,277 | 1,813 | -20% | 0 | 0 | — |
case-10 | fail→pass | 14,841 | 1,969 | -87% | 1 | 1 | 0% | 2,480 | 1,275 | -49% | 0 | 0 | — |
case-11 | fail→pass | 13,496 | 2,109 | -84% | 1 | 1 | 0% | 2,192 | 1,330 | -39% | 0 | 0 | — |
case-12 | fail→pass | 11,239 | 2,085 | -81% | 1 | 1 | 0% | 1,825 | 1,265 | -31% | 0 | 0 | — |
case-13 | fail→pass | 7,164 | 2,687 | -62% | 1 | 1 | 0% | 1,229 | 1,336 | +9% | 0 | 0 | — |
case-14 | pass→pass | 12,310 | 6,384 | -48% | 1 | 1 | 0% | 2,051 | 2,044 | -0% | 0 | 0 | — |
case-15 | pass→pass | 12,872 | 5,299 | -59% | 1 | 1 | 0% | 1,954 | 1,774 | -9% | 0 | 0 | — |
case-16 | fail→pass | 5,316 | 3,635 | -32% | 1 | 1 | 0% | 949 | 1,424 | +50% | 0 | 0 | — |
case-17 | fail→pass | 10,540 | 3,322 | -68% | 1 | 1 | 0% | 1,885 | 1,568 | -17% | 0 | 0 | — |
case-18 | fail→pass | 10,138 | 6,179 | -39% | 1 | 1 | 0% | 1,450 | 1,955 | +35% | 0 | 0 | — |
case-19 | fail→pass | 16,088 | 12,086 | -25% | 1 | 1 | 0% | 2,619 | 3,182 | +21% | 0 | 0 | — |
case-20 | pass→pass | 19,134 | 19,773 | +3% | 1 | 1 | 0% | 3,513 | 4,557 | +30% | 0 | 0 | — |
case-21 | pass→pass | 14,387 | 9,272 | -36% | 1 | 1 | 0% | 2,649 | 2,752 | +4% | 0 | 0 | — |
case-22 | pass→pass | 13,046 | 11,535 | -12% | 1 | 1 | 0% | 2,573 | 3,597 | +40% | 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 +59 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.