Agent skill · marketing growth · jeremylongshore
ideogram-incident-runbook
Analyze and contain Ideogram incidents involving credentials, billing, throttling, unsafe output, webhook loss, asset expiry, or storage failure. Use when coordinating production response and recovery. Trigger with "Ideogram incident", "contain an Ideogram outage", or "recover missing Ideogram images".
Why this skill is useful
Adds executable scripts for triaging and resolving Ideogram API incidents that aren't publicly documented.
What it needs
Requires bash, curl, kubectl installed locally. Requires ideogram account access. About 2k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Ideogram Incident Response Overview Stabilize an Ideogram integration before chasing image quality or replaying paid requests. Preserve evidence, reduce blast radius, reconcile accepted work, protect sensitive media, and choose recovery actions by incident class. Prerequisites Incident commander, environment, start time, impact statement, and communication channel. Read access to content-free application, queue, vendor-status, webhook, polling, storage, and billing evidence. Known controls for admission, concurrency, key revocation, traffic rollback, queue drain, and publication stop. Current Contract Incidents commonly cross separate boundaries: server-side Api-Key, prepaid shared team credit, default in-flight capacity, async generationid, signed but non-guaranteed webhook delivery, item safety, expiring URLs, and application storage. Recovery must not collapse these into a generic retry. Authentication Never paste the API key into incident chat or commands. Verify secret provenance through metadata; if exposure is credible, stop new traffic, revoke under owner approval, rotate consumers, and audit historical leakage. Instructions 1. Declare severity, impact, affected environment and tenants, known time window, and incident commander. 2. Freeze risky deployments and broad retries; preserve sanitized statuses, identifiers, queue state, and storage receipts. 3. Classify credential, depleted credit, validation, 429 pressure, vendor capacity, unsafe publication, webhook loss, expired URL, or storage failure. 4. Contain with the smallest control: close admission, reduce concurrency, stop publishing, switch to polling, isolate storage, or roll back traffic. 5. Reconcile every accepted async identifier before resubmission and prevent duplicate asset publication. 6. Recover with synthetic canaries, then restore bounded traffic while monitoring safety, durable completion, errors, latency, and spend. 7. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/ideogram-incident-runbook