Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Collect Clay debug evidence for support tickets and troubleshooting. Use when encountering persistent issues, preparing support tickets, or collecting diagnostic information for Clay integration problems. Trigger with phrases like "clay debug", "clay support bundle", "collect clay logs", "clay diagnostic", "clay support ticket".
.claude/skills/jeremylongshore-clay-debug-bundle/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 60% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 19% | 0% |
!node --version 2>/dev/null || echo 'N/A' !python3 --version 2>/dev/null || echo 'N/A'
Collect all diagnostic information needed for Clay support tickets. Clay is a web platform, so debugging focuses on webhook delivery, enrichment column errors, HTTP API responses, and credit consumption -- not pods or clusters.
bash#!/bin/bash # clay-debug-bundle.sh — collect Clay integration diagnostics set -euo pipefail BUNDLE_DIR="clay-debug-$(date +%Y%m%d-%H%M%S)" mkdir -p "$BUNDLE_DIR" echo "=== Clay Debug Bundle ===" > "$BUNDLE_DIR/summary.txt" echo "Generated: $(date -u +%Y-%m-%dT%H:%M:%SZ)" >> "$BUNDLE_DIR/summary.txt" echo "" >> "$BUNDLE_DIR/summary.txt"
bash# Test webhook endpoint is reachable echo "--- Webhook Test ---" >> "$BUNDLE_DIR/summary.txt" WEBHOOK_URL="${CLAY_WEBHOOK_URL:-not_set}" if [ "$WEBHOOK_URL" != "not_set" ]; then HTTP_CODE=$(curl -s -o "$BUNDLE_DIR/webhook-response.txt" -w "%{http_code}" \ -X POST "$WEBHOOK_URL" \ -H "Content-Type: application/json" \ -d '{"_debug": true, "_test": "debug-bundle"}') echo "Webhook HTTP Status: $HTTP_CODE" >> "$BUNDLE_DIR/summary.txt" echo "Webhook URL: ${WEBHOOK_URL:0:50}..." >> "$BUNDLE_DIR/summary.txt" else echo "CLAY_WEBHOOK_URL: NOT SET" >> "$BUNDLE_DIR/summary.txt" fi
bash# Capture environment (redacted) echo "" >> "$BUNDLE_DIR/summary.txt" echo "--- Environment ---" >> "$BUNDLE_DIR/summary.txt" echo "Node: $(node --version 2>/dev/null || echo 'N/A')" >> "$BUNDLE_DIR/summary.txt" echo "Python: $(python3 --version 2>/dev/null || echo 'N/A')" >> "$BUNDLE_DIR/summary.txt" echo "OS: $(uname -s -r)" >> "$BUNDLE_DIR/summary.txt" echo "CLAY_API_KEY: ${CLAY_API_KEY:+[SET]}" >> "$BUNDLE_DIR/summary.txt" echo "CLAY_WEBHOOK_URL: ${CLAY_WEBHOOK_URL:+[SET]}" >> "$BUNDLE_DIR/summary.txt" # Capture .env (redacted) if [ -f .env ]; then sed 's/=.*/=***REDACTED***/' .env > "$BUNDLE_DIR/config-redacted.txt" fi
bashecho "" >> "$BUNDLE_DIR/summary.txt" echo "--- Enterprise API Test ---" >> "$BUNDLE_DIR/summary.txt" if [ -n "${CLAY_API_KEY:-}" ]; then API_CODE=$(curl -s -o "$BUNDLE_DIR/api-response.txt" -w "%{http_code}" \ -X POST "https://api.clay.com/v1/people/enrich" \ -H "Authorization: Bearer $CLAY_API_KEY" \ -H "Content-Type: application/json" \ -d '{"email": "test@example.com"}') echo "Enterprise API Status: $API_CODE" >> "$BUNDLE_DIR/summary.txt" else echo "Enterprise API: No API key configured" >> "$BUNDLE_DIR/summary.txt" fi
bash# Grab recent Clay-related log entries echo "" >> "$BUNDLE_DIR/summary.txt" echo "--- Recent Errors ---" >> "$BUNDLE_DIR/summary.txt" # Search for Clay errors in common log locations for logfile in logs/*.log /tmp/clay-*.log; do if [ -f "$logfile" ]; then grep -i "clay\|webhook\|enrich" "$logfile" 2>/dev/null | tail -50 >> "$BUNDLE_DIR/logs.txt" fi done # Check npm logs for module issues if [ -d node_modules ]; then echo "Dependencies installed: YES" >> "$BUNDLE_DIR/summary.txt" ls package.json 2>/dev/null && echo "package.json: EXISTS" >> "$BUNDLE_DIR/summary.txt" fi
Add to the debug bundle manually from the Clay UI:
markdown## Clay Table Diagnostics (fill in from Clay UI) - Table Name: _______________ - Table URL: _______________ - Row Count: _______________ - Enrichment Columns: _______________ - Auto-run enabled: YES / NO - Error cells visible: YES (count: ___) / NO - Credit balance remaining: _______________ - Plan tier: Free / Starter / Explorer / Pro / Enterprise ## Error Details (from clicking red cells) - Error message: _______________ - Error column: _______________ - Error row count: _______________ ## Recent Changes - Last table edit: _______________ - Any new columns added: _______________ - Any provider connections changed: _______________
bashtar -czf "$BUNDLE_DIR.tar.gz" "$BUNDLE_DIR" echo "" echo "Bundle created: $BUNDLE_DIR.tar.gz" echo "" echo "BEFORE SUBMITTING: Review for sensitive data!" echo " - Check $BUNDLE_DIR/config-redacted.txt for leaked secrets" echo " - Check $BUNDLE_DIR/logs.txt for PII (emails, names)" echo "" echo "Submit at: https://community.clay.com or support@clay.com"
| Item | Purpose | Included | |------|---------|----------| | Webhook connectivity | Verify data can reach Clay | Automated | | Environment info | Compatibility check | Automated | | Enterprise API test | Verify API key validity | Automated | | Application logs | Error patterns | Automated | | Table diagnostics | Clay-side configuration | Manual checklist | | Config (redacted) | Secret management issues | Automated |
Always redact before submitting:
Safe to include:
For rate limit issues, see clay-rate-limits.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 22,574 | 15,301 | -32% | 1 | 1 | 0% | 4,646 | 4,699 | +1% | 0 | 0 | — |
case-02 | fail→pass | 21,679 | 15,634 | -28% | 1 | 1 | 0% | 4,327 | 4,968 | +15% | 0 | 0 | — |
case-03 | fail→fail | 21,519 | 16,505 | -23% | 1 | 1 | 0% | 3,966 | 5,235 | +32% | 0 | 0 | — |
case-04 | fail→pass | 26,031 | 6,939 | -73% | 1 | 1 | 0% | 1,608 | 2,812 | +75% | 0 | 0 | — |
case-05 | fail→pass | 19,758 | 7,439 | -62% | 1 | 1 | 0% | 1,973 | 3,152 | +60% | 0 | 0 | — |
case-06 | fail→pass | 10,094 | 2,405 | -76% | 1 | 1 | 0% | 1,726 | 2,148 | +24% | 0 | 0 | — |
case-07 | fail→pass | 16,702 | 9,811 | -41% | 1 | 1 | 0% | 2,879 | 3,427 | +19% | 0 | 0 | — |
case-08 | fail→pass | 13,172 | 10,192 | -23% | 1 | 1 | 0% | 1,964 | 3,159 | +61% | 0 | 0 | — |
case-09 | fail→fail | 10,255 | 5,075 | -51% | 1 | 1 | 0% | 1,816 | 2,599 | +43% | 0 | 0 | — |
case-21 | pass→pass | 14,410 | 13,943 | -3% | 1 | 1 | 0% | 2,656 | 4,601 | +73% | 0 | 0 | — |
case-10 | fail→pass | 12,001 | 5,498 | -54% | 1 | 1 | 0% | 2,253 | 2,800 | +24% | 0 | 0 | — |
case-11 | fail→pass | 10,770 | 4,088 | -62% | 1 | 1 | 0% | 1,678 | 2,283 | +36% | 0 | 0 | — |
case-12 | fail→pass | 14,012 | 3,219 | -77% | 1 | 1 | 0% | 2,174 | 2,167 | -0% | 0 | 0 | — |
case-13 | fail→pass | 13,049 | 8,820 | -32% | 1 | 1 | 0% | 1,912 | 3,065 | +60% | 0 | 0 | — |
case-14 | fail→pass | 8,158 | 2,224 | -73% | 1 | 1 | 0% | 1,051 | 2,037 | +94% | 0 | 0 | — |
case-15 | pass→pass | 12,642 | 7,759 | -39% | 1 | 1 | 0% | 2,196 | 3,078 | +40% | 0 | 0 | — |
case-16 | fail→fail | 10,609 | 2,288 | -78% | 1 | 1 | 0% | 1,644 | 2,074 | +26% | 0 | 0 | — |
case-17 | fail→pass | 12,542 | 8,473 | -32% | 1 | 1 | 0% | 1,984 | 3,183 | +60% | 0 | 0 | — |
case-18 | fail→pass | 11,901 | 6,303 | -47% | 1 | 1 | 0% | 1,817 | 2,631 | +45% | 0 | 0 | — |
case-19 | fail→pass | 11,499 | 3,311 | -71% | 1 | 1 | 0% | 2,066 | 2,209 | +7% | 0 | 0 | — |
case-20 | pass→pass | 13,406 | 10,680 | -20% | 1 | 1 | 0% | 2,698 | 3,629 | +35% | 0 | 0 | — |
case-22 | pass→pass | 11,890 | 9,219 | -22% | 1 | 1 | 0% | 2,119 | 3,262 | +54% | 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, and 21 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 +64 percentage points is the difference between those two pass rates over the 21 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.