Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Collect debug evidence for Anthropic API issues — request IDs, headers, Use when working with debug-bundle patterns. error payloads, and reproduction steps for support tickets. Trigger with "anthropic debug", "claude support ticket", "anthropic request id", "debug claude api call".
.claude/skills/jeremylongshore-clade-debug-bundle/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 132% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 109% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 23% | 0% |
When you need to file a support ticket or debug a persistent issue, collect these items.
Every Anthropic API response includes a request-id header. This is the single most important thing for support tickets.
typescripttry { const message = await client.messages.create({ ... }); // Access response headers via the raw response } catch (err) { if (err instanceof Anthropic.APIError) { console.error('Request ID:', err.headers?.['request-id']); console.error('Status:', err.status); console.error('Error type:', err.error?.type); console.error('Message:', err.message); } }
typescriptfunction logAnthropicError(err: unknown) { if (err instanceof Anthropic.APIError) { const bundle = { timestamp: new Date().toISOString(), request_id: err.headers?.['request-id'], status: err.status, error_type: err.error?.type, error_message: err.message, rate_limit_remaining: err.headers?.['claude-ratelimit-requests-remaining'], rate_limit_reset: err.headers?.['claude-ratelimit-requests-reset'], }; console.error('Anthropic Debug Bundle:', JSON.stringify(bundle, null, 2)); return bundle; } console.error('Non-API error:', err); }
bash# Minimal reproduction — include this in support tickets curl -v https://api.anthropic.com/v1/messages \ -H "x-api-key: $ANTHROPIC_API_KEY" \ -H "claude-version: 2023-06-01" \ -H "content-type: application/json" \ -d '{ "model": "claude-sonnet-4-20250514", "max_tokens": 100, "messages": [{"role": "user", "content": "test"}] }' 2>&1 | grep -E "request-id|HTTP|error"
bash# API status curl -s https://status.anthropic.com/api/v2/status.json | python3 -m json.tool # Recent incidents curl -s https://status.anthropic.com/api/v2/incidents.json | python3 -c " import json, sys data = json.load(sys.stdin) for inc in data['incidents'][:3]: print(f\"{inc['created_at'][:10]}: {inc['name']} ({inc['status']})\") "
request-id header)npm list @claude-ai/sdk or pip show anthropic)pythontry: message = client.messages.create(...) except anthropic.APIStatusError as e: print(f"Request ID: {e.response.headers.get('request-id')}") print(f"Status: {e.status_code}") print(f"Error: {e.message}")
| Error | Cause | Solution | |-------|-------|----------| | API Error | Check error type and status code | See clade-common-errors |
See Step 1 (request ID extraction), Step 2 (full error logging), Step 3 (curl reproduction), and Step 4 (status check) above.
See clade-common-errors for specific error solutions.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 18,151 | 13,669 | -25% | 1 | 1 | 0% | 3,865 | 4,088 | +6% | 0 | 0 | — |
case-02 | fail→pass | 16,922 | 9,779 | -42% | 1 | 1 | 0% | 3,320 | 2,999 | -10% | 0 | 0 | — |
case-03 | pass→fail | 13,605 | 11,982 | -12% | 1 | 1 | 0% | 2,563 | 3,617 | +41% | 0 | 0 | — |
case-04 | pass→pass | 7,865 | 6,777 | -14% | 1 | 1 | 0% | 1,472 | 2,545 | +73% | 0 | 0 | — |
case-05 | pass→pass | 7,984 | 2,182 | -73% | 1 | 1 | 0% | 1,365 | 1,432 | +5% | 0 | 0 | — |
case-06 | fail→pass | 10,553 | 7,436 | -30% | 1 | 1 | 0% | 1,645 | 2,413 | +47% | 0 | 0 | — |
case-07 | fail→pass | 4,933 | 3,212 | -35% | 1 | 1 | 0% | 740 | 1,719 | +132% | 0 | 0 | — |
case-08 | fail→pass | 4,632 | 4,532 | -2% | 1 | 1 | 0% | 875 | 1,828 | +109% | 0 | 0 | — |
case-09 | fail→fail | 6,566 | 3,099 | -53% | 1 | 1 | 0% | 516 | 1,647 | +219% | 0 | 0 | — |
case-10 | pass→pass | 21,465 | 9,495 | -56% | 1 | 1 | 0% | 2,133 | 2,795 | +31% | 0 | 0 | — |
case-11 | pass→fail | 3,716 | 1,757 | -53% | 1 | 1 | 0% | 657 | 1,273 | +94% | 0 | 0 | — |
case-12 | pass→pass | 3,486 | 2,094 | -40% | 1 | 1 | 0% | 602 | 1,317 | +119% | 0 | 0 | — |
case-13 | pass→pass | 3,880 | 3,556 | -8% | 1 | 1 | 0% | 587 | 1,591 | +171% | 0 | 0 | — |
case-14 | pass→pass | 9,261 | 6,227 | -33% | 1 | 1 | 0% | 1,680 | 2,350 | +40% | 0 | 0 | — |
case-15 | pass→pass | 10,028 | 7,953 | -21% | 1 | 1 | 0% | 1,738 | 1,873 | +8% | 0 | 0 | — |
case-16 | fail→fail | 7,748 | 6,042 | -22% | 1 | 1 | 0% | 1,479 | 2,163 | +46% | 0 | 0 | — |
case-17 | pass→pass | 11,937 | 2,773 | -77% | 1 | 1 | 0% | 2,032 | 1,618 | -20% | 0 | 0 | — |
case-18 | fail→pass | 7,517 | 2,270 | -70% | 1 | 1 | 0% | 1,231 | 1,516 | +23% | 0 | 0 | — |
case-19 | pass→pass | 3,289 | 1,918 | -42% | 1 | 1 | 0% | 505 | 1,402 | +178% | 0 | 0 | — |
case-20 | pass→pass | 13,691 | 13,965 | +2% | 1 | 1 | 0% | 2,624 | 3,948 | +50% | 0 | 0 | — |
case-21 | pass→pass | 10,048 | 10,188 | +1% | 1 | 1 | 0% | 1,965 | 2,948 | +50% | 0 | 0 | — |
case-22 | pass→pass | 14,793 | 19,688 | +33% | 1 | 1 | 0% | 2,726 | 4,927 | +81% | 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 +14 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.