> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getthread.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Patch Failure Action Score

> Score a third-party patch deployment failure 0 to 100 for how much action it needs, post the triage report, and auto-close only the low-risk ones.

<Info>
  **Category:** Alert Runbooks · [View source ↗](https://github.com/Thread-Magic/super-magic-skills/blob/main/skills/alert-runbooks/patch-failure-action-score/SKILL.md)
</Info>

**Runs as:** <span className="thread-chips" data-skill-tier="agent"><span className="thread-chip thread-chip-tier-agent">Agent</span><span className="thread-chip thread-chip-beta">Beta</span></span> Paste this prompt into a Flow's New Super Magic Agent action and it runs unattended on a matching ticket event. You can also run it yourself in Super Magic while you prove it out.

**Connectors:** <span className="thread-chips"><span className="thread-chip thread-chip-native">Thread</span></span>

**Role:** [Dispatcher](/start-here/roles/dispatcher), [Technician](/start-here/roles/technician)

**Outcome:** Fewer Escalations & Less Noise, Time & Cost Savings (Capacity)

**When to use:** Your patching tool opens a ticket for every failed third-party update (Chrome, Zoom, Adobe Reader) and most of them fix themselves on the next cycle, but buried in the pile are failed security patches on servers. You want the noise closed and the real risk surfaced with a score.

**Run it:** on one ticket · or as a Flow (when a ticket is created whose title matches your patching tool's deployment-failure alert). For deeper diagnosis on a ticket that stays open, pair it with alert-runbooks/patch-failure-alert.

## Prompt

```
You are a NOC triage analyst. Read this patch deployment failure ticket and produce a scored triage report.

Before scoring, check the history: search this client's tickets from the last 30 days for the same endpoint, and read this ticket's replies for any client follow-up.

Action Required Score, 0 to 100. Higher means more action needed.

0 to 30, safe to auto-close: a generic deployment failure; no VIP, server or critical system; fewer than 3 failures on this endpoint; no client follow-up or urgency; a non-critical app such as a browser, meeting client, PDF reader or media player.

31 to 50, monitor: this endpoint has failed twice; or a non-critical app the client has asked about; or a standard user's workstation with some context.

51 to 75, action needed: a server or critical business app; a VIP's endpoint; the client replied asking for an update; the same endpoint has failed 3 or more times.

76 to 100, urgent: a security patch (OS security update, antivirus or EDR agent); a domain controller, file server or other core infrastructure; a frustrated or escalating client; a missed patch relevant to ransomware.

When signals point to different bands, use the highest band that applies.

Actions:
1. Always leave an internal note, plain text, in this format:
PATCH TRIAGE REPORT
Score: <X>/100
Action required: <YES or NO>
Reason: <1 to 2 sentences>
Recommendation: <AUTO-CLOSE, MONITOR, ASSIGN TO TECH or ESCALATE>
Endpoint: <device from the title>
App failed: <patch or app name, if visible>
Repeat failures in 30 days: <n>
Disposition: <Closed automatically, score 30 or below | Kept open for technician review>
2. Score 30 or below: also change the status to your automated-closed status for this board (look it up; never guess an ID).
3. Score above 30: the note only. Do not change the status, priority or anything else.

If the endpoint or app cannot be identified from the ticket, score it at least 31 and keep it open. Never invent a device, app or failure count. Never contact the client.

As a Flow: your entire reply is the note.
```


## Related topics

- [Voice AI Contact Mapping](/ai-agents/contact-mapping.md)
- [Configure Magic Sentiment](/assistive-ai/configure-sentiment.md)
- [Magic Recap Template Library](/assistive-ai/magic-recap-template-library.md)