Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Respond to Figma API outages, auth failures, and rate limit incidents. Use when Figma integration is down, experiencing errors, or running post-incident reviews for Figma-related failures. Trigger with phrases like "figma incident", "figma outage", "figma down", "figma broken", "figma emergency".
.claude/skills/jeremylongshore-figma-incident-runbook/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 120% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 78% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -13% | 0% |
Rapid incident response procedures for Figma REST API integration failures. Covers triage, mitigation, and postmortem for the most common failure modes.
bash#!/bin/bash echo "=== Figma Incident Triage ===" # 1. Is Figma itself down? echo -n "Figma Status: " curl -s https://www.figmastatus.com/api/v2/status.json 2>/dev/null \ | jq -r '.status.description // "Cannot reach status page"' # 2. Is our token valid? echo -n "Auth Check: " HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" \ -H "X-Figma-Token: ${FIGMA_PAT}" \ https://api.figma.com/v1/me) echo "$HTTP_CODE" # 3. Can we read a known file? echo -n "File Access: " curl -s -H "X-Figma-Token: ${FIGMA_PAT}" \ "https://api.figma.com/v1/files/${FIGMA_FILE_KEY}?depth=1" \ | jq -r '.name // "FAILED"' # 4. Are we rate limited? echo "Rate Limit Headers:" curl -s -D - -o /dev/null \ -H "X-Figma-Token: ${FIGMA_PAT}" \ https://api.figma.com/v1/me 2>/dev/null \ | grep -iE "(retry-after|rate-limit|figma)" || echo "No rate limit headers"
API returning errors?
├── 403 Forbidden
│ ├── Token expired (>90 days) → Rotate PAT immediately
│ ├── Wrong scopes → Regenerate with correct scopes
│ └── File not shared → Check file permissions
│
├── 429 Rate Limited
│ ├── Retry-After < 60s → Wait and retry automatically
│ ├── Retry-After > 300s → Reduce request volume
│ └── X-Figma-Rate-Limit-Type: low → Consider upgrading plan
│
├── 404 Not Found
│ ├── File deleted → Check with file owner
│ ├── Wrong file key → Verify FIGMA_FILE_KEY
│ └── API path wrong → Check endpoint documentation
│
├── 500/503 Server Error
│ ├── status.figma.com shows incident → Wait for resolution
│ ├── Intermittent → Retry with backoff
│ └── Persistent → Contact Figma support
│
└── Network Error (ECONNREFUSED, timeout)
├── DNS resolution failing → Check DNS config
├── Firewall blocking → Verify outbound HTTPS to api.figma.com
└── TLS error → Check Node.js version (18+ required)For 403 (Token Expired):
bash# Generate new PAT in Figma Settings > Personal access tokens # Then update your deployment: # GitHub Actions gh secret set FIGMA_PAT --body "figd_new-token-here" # Cloud Run echo -n "figd_new-token" | gcloud secrets versions add figma-pat --data-file=- gcloud run services update my-service --update-secrets="FIGMA_PAT=figma-pat:latest" # Fly.io fly secrets set FIGMA_PAT=figd_new-token
For 429 (Rate Limited):
typescript// Emergency: disable non-critical Figma calls const EMERGENCY_MODE = process.env.FIGMA_EMERGENCY === 'true'; async function safeFigmaCall<T>( path: string, critical: boolean = false ): Promise<T | null> { if (EMERGENCY_MODE && !critical) { console.warn(`Figma call skipped (emergency mode): ${path}`); return null; } return figmaFetch(path); }
For 500/503 (Figma Down):
typescript// Serve cached data when Figma is unavailable async function getTokensWithFallback() { try { return await extractTokensFromFigma(); } catch (error) { console.warn('Figma unavailable, serving cached tokens'); // Return last-known-good tokens from cache or file const cached = await readFile('output/tokens.json', 'utf-8'); return JSON.parse(cached); } }
markdown## Internal Notification (Slack) **Figma Integration Alert** - Status: INVESTIGATING / MITIGATED / RESOLVED - Impact: [Design token sync paused / Asset export failing] - Cause: [403 expired token / 429 rate limit / Figma outage] - Action: [Rotating token / Reducing request rate / Waiting for Figma] - ETA: [Next update in 15 min] ## External (if applicable) Design system updates may be delayed due to a temporary issue with our Figma integration. Cached data is being served.
markdown## Figma Incident Postmortem **Date:** YYYY-MM-DD **Duration:** X hours Y minutes **Severity:** P1/P2/P3 ### Summary [One sentence: what happened and what was the impact] ### Timeline - HH:MM UTC - First alert fired (describe alert) - HH:MM UTC - On-call acknowledged - HH:MM UTC - Root cause identified - HH:MM UTC - Mitigation applied - HH:MM UTC - Full resolution confirmed ### Root Cause [Technical explanation, e.g., "PAT expired after 90 days without rotation"] ### Action Items - [ ] Set up PAT rotation reminder at 80-day mark - [ ] Add 403 alert to PagerDuty - [ ] Implement cached fallback for token data
| Issue | Cause | Solution | |-------|-------|----------| | Can't reach status.figma.com | Network issue | Try from different network or mobile | | Triage script fails | PAT not set | Set FIGMA_PAT before running | | Fallback data stale | Last cache too old | Set up regular cache refresh | | Alert not firing | Missing metrics | Verify Prometheus scrape config |
Triage "the design token sync is failing" in under a minute (Step 1):
bash# Is it us or Figma? Probe auth, then a known-good file with minimal payload curl -s -o /dev/null -w '%{http_code}\n' -H "X-Figma-Token: ${FIGMA_PAT}" https://api.figma.com/v1/me curl -s -o /dev/null -w '%{http_code}\n' -H "X-Figma-Token: ${FIGMA_PAT}" \ "https://api.figma.com/v1/files/${FIGMA_FILE_KEY}?depth=1"
| /v1/me | file probe | Read as | |----------|-----------|---------| | 200 | 200 | Our pipeline bug — not an API incident | | 200 | 403/404 | File access changed — check sharing/scopes | | 401 | any | Token revoked or expired — rotate now | | 429 / 5xx | 429 / 5xx | Figma-side — mitigate per Step 3, check status.figma.com |
Decision tree and mitigation actions: references/decision-tree.md, references/immediate-mitigation.md; postmortem scaffold: references/postmortem-template.md.
For data handling, see figma-data-handling.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 26,508 | 19,453 | -27% | 1 | 1 | 0% | 5,430 | 5,412 | -0% | 0 | 0 | — |
case-02 | fail→fail | 22,942 | 17,338 | -24% | 1 | 1 | 0% | 3,785 | 5,585 | +48% | 0 | 0 | — |
case-03 | fail→fail | 19,824 | 21,228 | +7% | 1 | 1 | 0% | 3,222 | 5,772 | +79% | 0 | 0 | — |
case-04 | pass→pass | 21,399 | 23,205 | +8% | 1 | 1 | 0% | 4,335 | 5,601 | +29% | 0 | 0 | — |
case-05 | fail→fail | 15,053 | 14,566 | -3% | 1 | 1 | 0% | 2,910 | 4,895 | +68% | 0 | 0 | — |
case-06 | pass→pass | 16,948 | 18,238 | +8% | 1 | 1 | 0% | 3,308 | 5,894 | +78% | 0 | 0 | — |
case-07 | pass→pass | 12,843 | 9,045 | -30% | 1 | 1 | 0% | 2,451 | 3,896 | +59% | 0 | 0 | — |
case-08 | pass→pass | 9,239 | 7,217 | -22% | 1 | 1 | 0% | 1,643 | 3,101 | +89% | 0 | 0 | — |
case-09 | pass→pass | 5,163 | 4,957 | -4% | 1 | 1 | 0% | 1,018 | 2,882 | +183% | 0 | 0 | — |
case-10 | fail→pass | 133,273 | 15,497 | -88% | 1 | 1 | 0% | 3,331 | 5,104 | +53% | 0 | 0 | — |
case-11 | pass→pass | 19,541 | 10,161 | -48% | 1 | 1 | 0% | 2,738 | 3,467 | +27% | 0 | 0 | — |
case-12 | fail→pass | 12,916 | 16,654 | +29% | 1 | 1 | 0% | 2,185 | 4,806 | +120% | 0 | 0 | — |
case-13 | pass→pass | 5,800 | 5,445 | -6% | 1 | 1 | 0% | 1,103 | 2,980 | +170% | 0 | 0 | — |
case-14 | fail→fail | 8,579 | 4,747 | -45% | 1 | 1 | 0% | 1,634 | 2,984 | +83% | 0 | 0 | — |
case-15 | fail→pass | 11,531 | 8,954 | -22% | 1 | 1 | 0% | 1,932 | 3,447 | +78% | 0 | 0 | — |
case-16 | fail→pass | 15,227 | 6,596 | -57% | 1 | 1 | 0% | 2,415 | 3,170 | +31% | 0 | 0 | — |
case-17 | fail→pass | 17,779 | 3,835 | -78% | 1 | 1 | 0% | 3,056 | 2,672 | -13% | 0 | 0 | — |
case-18 | pass→pass | 16,066 | 12,279 | -24% | 1 | 1 | 0% | 2,547 | 3,988 | +57% | 0 | 0 | — |
case-19 | pass→pass | 8,021 | 7,703 | -4% | 1 | 1 | 0% | 1,268 | 3,187 | +151% | 0 | 0 | — |
case-20 | pass→pass | 15,648 | 11,513 | -26% | 1 | 1 | 0% | 2,271 | 3,693 | +63% | 0 | 0 | — |
case-21 | pass→pass | 8,848 | 6,220 | -30% | 1 | 1 | 0% | 1,599 | 3,146 | +97% | 0 | 0 | — |
case-22 | fail→fail | 10,551 | 8,221 | -22% | 1 | 1 | 0% | 2,077 | 3,596 | +73% | 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 +23 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.