LLM Incident Response Runbook System
When your LLM feature breaks in production, stop improvising — classify the incident into one of four pre-built classes, assign a defensible severity, walk a class-specific diagnosis tree to the root cause, ship a sequenced containment plan, send the right comms, and close it with a blameless postmortem.
A 5-step agentic workflow pack for coding built to run with ChatGPT, Claude, and Gemini. Open the Markdown files, fill the variables, and paste into your model. Most buyers get a reviewable result in about 15 minutes.
- Classify any live LLM incident into one of four pre-built classes — quality regression, cost/latency spike, prompt-injection event, or data-leak-via-output — instead of debating what is even happening
- Assign a defensible SEV1-SEV4 severity from a weighted matrix (blast radius, data exposure, trust impact, reversibility) so you page the right people without over- or under-reacting
- Walk a class-specific diagnosis decision tree with a named failure-mode catalog for each branch — isolate the broken layer instead of guessing between the model, the prompt, retrieval, and the change log
- Get a sequenced containment plan that decides rollback vs mitigate vs kill-switch, time-boxes each step, and includes verification checks that prove the bleeding stopped
- Send the right message to the right audience with internal status updates, a customer-facing statement, and an exec brief calibrated to severity and cadence
- Close every incident with a blameless postmortem whose action items map to LLM-specific preventive controls — eval gates, canary rollout, cost guardrails, injection filters, and output DLP
- Encodes practitioner sequencing and scoring judgment, not a generic 'ask the model what to do' — the runbook is the same every time so on-call engineers act consistently under pressure
Prompt Customization Service — optional help adapting variables and output to your brand voice. Choose your tier at checkout (not tied to this prompt's price).
This pack is $10 on its own. Buying every pack separately costs $1096. The Lifetime Bundle is $149 one-time — you save $947 (86% off) and unlock every future pack free.
Get the Lifetime Bundle — $149Paste the license key from your receipt. It must match this prompt pack.
What ships with your purchase
Prompt files
Plain Markdown files with `{{variables}}` you fill in, ready to paste into ChatGPT, Claude, or Gemini. No setup, no tooling required.
Usage guide
Variable reference, model compatibility, examples, and customization tips so you can adapt the pack to your brand voice.
Lifetime updates
When we improve the pack, you get the new version automatically. Email support included with every purchase.
Models tested: ChatGPT, Claude, Gemini.
The workflow inside this pack
5 composable prompts you run in order — each one picks up where the last left off.
- Step 1
Incident Triage & Severity Classifier
Paste the incident symptoms, the affected surface, and your recent deployment change log — each incident is matched to one of four pre-built classes with the evidence behind the match.
- Step 2
Diagnosis Tree Navigator
Feed in the incident class and severity from the triage step plus your logs, traces, and metric deltas — the prompt walks the decision tree built for that exact class, branch by branch.
- Step 3
Containment & Mitigation Planner
Give it the incident class, the leading root-cause hypothesis, and your available controls — rollback targets, feature flags, rate limits, and kill switches — and it weighs rollback against forward mitigation.
- Step 4 · optional
Stakeholder Comms Composer
Paste the severity and class, the current status from the containment plan, and who needs to hear from you — the prompt writes each message tuned to that audience and severity.
- Step 5 · optional
Blameless Postmortem Builder
Give it the timeline, the confirmed root cause, the containment actions taken, and the measured impact — it assembles a blameless narrative with contributing factors, not a single culprit.
Perpetual (lifetime) use license
Your one-time purchase includes an ongoing right to use this prompt pack with the AI tools and models you control for your own and your clients' work — not for resale or public redistribution of the files as a product.
We keep the copyright
The prompt files, guides, examples, and bundled assets stay our copyrighted works (or our licensors'). Payment grants the limited license in our Terms only — it does not transfer ownership.
Need help adapting this prompt to your team? Add Prompt Customization Service at checkout.