Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Expert patterns for Segment Customer Data Platform including Analytics.js, server-side tracking, tracking plans with Protocols, identity resolution, destinations configuration, and data governance best practices. Use when: segment, analytics.js, customer data platform, cdp, tracking plan.
.claude/skills/davila7-segment-cdp/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -17% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -15% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -14% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -8% | 0% |
Client-side tracking with Analytics.js. Include track, identify, page, and group calls. Anonymous ID persists until identify merges with user.
High-performance server-side tracking using @segment/analytics-node. Non-blocking with internal batching. Essential for backend events, webhooks, and sensitive data.
Design event schemas using Object + Action naming convention. Define required properties, types, and validation rules. Connect to Protocols for enforcement.
| Issue | Severity | Solution | |-------|----------|----------| | Issue | medium | See docs | | Issue | high | See docs | | Issue | medium | See docs | | Issue | high | See docs | | Issue | low | See docs | | Issue | medium | See docs | | Issue | medium | See docs | | Issue | high | See docs |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | pass→pass | 11,654 | 8,049 | -31% | 1 | 1 | 0% | 1,946 | 1,612 | -17% | 0 | 0 | — |
case-01 | fail→pass | 16,048 | 14,157 | -12% | 1 | 1 | 0% | 3,051 | 2,792 | -8% | 0 | 0 | — |
case-03 | pass→pass | 10,763 | 7,840 | -27% | 1 | 1 | 0% | 1,852 | 1,568 | -15% | 0 | 0 | — |
case-04 | pass→pass | 10,979 | 7,696 | -30% | 1 | 1 | 0% | 1,857 | 1,590 | -14% | 0 | 0 | — |
case-05 | pass→pass | 13,540 | 10,851 | -20% | 1 | 1 | 0% | 2,196 | 2,030 | -8% | 0 | 0 | — |
case-06 | pass→pass | 16,035 | 12,328 | -23% | 1 | 1 | 0% | 2,723 | 2,513 | -8% | 0 | 0 | — |
case-07 | pass→pass | 13,744 | 10,522 | -23% | 1 | 1 | 0% | 2,353 | 2,185 | -7% | 0 | 0 | — |
case-08 | pass→pass | 14,978 | 10,694 | -29% | 1 | 1 | 0% | 2,355 | 2,015 | -14% | 0 | 0 | — |
case-09 | pass→pass | 5,569 | 8,888 | +60% | 1 | 1 | 0% | 866 | 1,635 | +89% | 0 | 0 | — |
case-10 | pass→pass | 19,182 | 13,848 | -28% | 1 | 1 | 0% | 3,625 | 3,054 | -16% | 0 | 0 | — |
case-11 | pass→pass | 13,707 | 10,389 | -24% | 1 | 1 | 0% | 2,298 | 2,012 | -12% | 0 | 0 | — |
case-12 | pass→pass | 11,799 | 6,863 | -42% | 1 | 1 | 0% | 1,847 | 1,286 | -30% | 0 | 0 | — |
case-22 | pass→pass | 10,923 | 10,233 | -6% | 1 | 1 | 0% | 2,078 | 2,131 | +3% | 0 | 0 | — |
case-13 | pass→pass | 16,156 | 13,101 | -19% | 1 | 1 | 0% | 2,763 | 2,383 | -14% | 0 | 0 | — |
case-14 | pass→pass | 13,455 | 10,005 | -26% | 1 | 1 | 0% | 2,281 | 1,998 | -12% | 0 | 0 | — |
case-15 | pass→pass | 6,644 | 4,652 | -30% | 1 | 1 | 0% | 1,013 | 968 | -4% | 0 | 0 | — |
case-16 | pass→pass | 11,062 | 6,937 | -37% | 1 | 1 | 0% | 1,813 | 1,389 | -23% | 0 | 0 | — |
case-17 | pass→pass | 11,705 | 10,808 | -8% | 1 | 1 | 0% | 1,993 | 2,032 | +2% | 0 | 0 | — |
case-18 | pass→pass | 10,335 | 8,633 | -16% | 1 | 1 | 0% | 1,950 | 1,846 | -5% | 0 | 0 | — |
case-19 | pass→pass | 15,176 | 13,953 | -8% | 1 | 1 | 0% | 2,698 | 2,899 | +7% | 0 | 0 | — |
case-20 | pass→pass | 15,274 | 13,497 | -12% | 1 | 1 | 0% | 2,426 | 2,539 | +5% | 0 | 0 | — |
case-21 | pass→pass | 9,466 | 8,458 | -11% | 1 | 1 | 0% | 1,522 | 1,588 | +4% | 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 +5 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.