Imported from jeremylongshore/tons-of-skills-marketplace (skills/.curated/fireflies-incident-runbook/SKILL.md). Install upstream with npx skills add jeremylongshore/tons-of-skills-marketplace --skill fireflies-incident-runbook. Copyright stays with the author (MIT).

Fireflies.ai Incident Runbook

Overview

Rapid incident response procedures for Fireflies.ai integration failures. Covers API outages, authentication problems, webhook issues, and rate limiting.

Prerequisites

• A declared incident commander, communication owner, secure evidence location, and escalation path.
• Credential-revocation and queue-disable procedures, plus redaction rules for transcripts and participants.
• Aggregate health probes that can be run without retrieving a real transcript.

Instructions

1. Assign severity and incident owner, record an opaque incident ID, and pause unsafe automation or sharing paths.
2. Classify the event as availability, access, webhook, throttle, retention, or potential data exposure.
3. Apply the smallest safe mitigation, verify recovery with a synthetic/read-only probe, and test the safe failure path.
4. Revoke or rotate credentials when exposure is possible; do not resume queued work until permission and idempotency checks pass.
5. Record timeline, impact, owner decisions, corrective actions, and the post-incident review date.

Examples

During a synthetic API outage, pause a transcript-routing worker, verify that queued events do not resend, and publish a redacted status receipt. Restore one canary only after the health probe recovers, then resume production processing with commander approval.

Severity Levels

Quick Triage (Run First)

set -euo pipefail
echo "=== Fireflies.ai Incident Triage ==="
echo ""

# 1. Can we reach the API?
echo "--- API Connectivity ---"
curl -s -o /dev/null -w "HTTP %{http_code} (%{time_total}s)\n" \
  -X POST https://api.fireflies.ai/graphql \
  -H "Authorization: Bearer $FIREFLIES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "{ user { email } }"}'

# 2. What error are we getting?
echo ""
echo "--- API Response ---"
curl -s -X POST https://api.fireflies.ai/graphql \
  -H "Authorization: Bearer $FIREFLIES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "{ user { email is_admin } }"}' | jq .

# 3. Can we list transcripts?
echo ""
echo "--- Transcript Access ---"
curl -s -X POST https://api.fireflies.ai/graphql \
  -H "Authorization: Bearer $FIREFLIES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "{ transcripts(limit: 1) { id title date } }"}' | jq '.data.transcripts[0] // .errors[0]'

Decision Tree

API returning errors?
├─ YES: What error code?
│   ├─ auth_failed (401) → API key revoked or invalid
│   │   └─ Fix: Regenerate key at app.fireflies.ai > Integrations
│   ├─ too_many_requests (429) → Rate limited
│   │   └─ Fix: Enable backoff, check for runaway loops
│   ├─ account_cancelled (403) → Subscription expired
│   │   └─ Fix: Renew subscription, contact billing
│   └─ 5xx errors → Fireflies platform issue
│       └─ Fix: Enable fallback mode, wait for resolution
└─ NO: Webhook issues?
    ├─ Not receiving webhooks → Check dashboard registration
    ├─ Invalid signature → Webhook secret mismatch
    └─ Processing failures → Check your webhook handler logs

Remediation by Error Type

auth_failed (401) -- P1

set -euo pipefail
# Verify API key format (should be non-empty)
echo "Key length: ${#FIREFLIES_API_KEY}"

# Test with explicit key
curl -s -X POST https://api.fireflies.ai/graphql \
  -H "Authorization: Bearer $FIREFLIES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "{ user { email } }"}' | jq '.errors[0].code // "OK"'

# Fix: Regenerate at app.fireflies.ai > Integrations > Fireflies API
# Then update your secret store:
# - GitHub: gh secret set FIREFLIES_API_KEY --body "new-key"
# - Vercel: vercel env rm FIREFLIES_API_KEY && vercel env add FIREFLIES_API_KEY
# - GCP: echo -n "new-key" | gcloud secrets versions add fireflies-api-key --data-file=-

too_many_requests (429) -- P2

set -euo pipefail
# Check if we have a request loop
# Free/Pro: 50/day limit. Business: 60/min limit.
echo "Check application logs for request volume"

# Immediate mitigation: reduce request rate
# Long-term: implement exponential backoff (see fireflies-rate-limits skill)

Webhook Not Firing -- P2

set -euo pipefail
# Verify webhook is registered
echo "Check: app.fireflies.ai > Settings > Developer settings"
echo "Webhook URL should be your HTTPS endpoint"

# Test by uploading audio (triggers webhook when done)
curl -s -X POST https://api.fireflies.ai/graphql \
  -H "Authorization: Bearer $FIREFLIES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "mutation($input: AudioUploadInput) { uploadAudio(input: $input) { success message } }",
    "variables": { "input": { "url": "https://example.com/test.mp3", "title": "Webhook Test" } }
  }' | jq .

# Remember: webhooks only fire for meetings YOU own (organizer_email)

Invalid Webhook Signature -- P3

// Debug signature verification
import crypto from "crypto";

function debugWebhookSignature(payload: string, receivedSig: string, secret: string) {
  const computed = crypto.createHmac("sha256", secret).update(payload).digest("hex");

  console.log("Received signature:", receivedSig);
  console.log("Computed signature:", computed);
  console.log("Match:", receivedSig === computed);
  console.log("Secret length:", secret.length, "(must be 16-32)");

  // Common issues:
  // 1. Secret doesn't match what's in Fireflies dashboard
  // 2. Payload was parsed/modified before verification (use raw body)
  // 3. Secret has trailing whitespace
}

Communication Templates

Internal (Slack)

P[1-4] INCIDENT: Fireflies.ai Integration
Status: INVESTIGATING / MITIGATED / RESOLVED
Error: [error code and message]
Impact: [what users are affected]
Action: [what we're doing]
ETA: [next update time]

Postmortem Template

## Incident: Fireflies.ai [Error Type]
**Date:** YYYY-MM-DD | **Duration:** Xh Ym | **Severity:** P[1-4]

### Summary
[1-2 sentences]

### Timeline
- HH:MM - Issue detected via [alert/user report]
- HH:MM - Triage started
- HH:MM - Root cause identified: [cause]
- HH:MM - Fix applied
- HH:MM - Verified resolved

### Root Cause
[Technical explanation]

### Action Items
- [ ] [Prevention measure] - Owner - Due date

Error Handling

Output

• Issue identified and categorized by severity
• Remediation applied based on error code
• Stakeholders notified via templates
• Evidence collected for postmortem

Resources

• Fireflies API Docs
• Fireflies Webhooks

Next Steps

For data handling and compliance, see fireflies-data-handling.