Skills Compliance
Scanning for harm you have no category for
Every monitoring system in a support operation looks for things someone already thought of. Tags, alerts, complaint categories and QA criteria all encode last year's known failures.
npx rulebase-skills install cx-emerging-harm-scanWhen to use it
Reach for this when someone says any of these — they are the phrases the skill itself triggers on:
- “is anything new going wrong”
- “scan for emerging issues”
- “early warning from support”
How it works
The method, in the order the skill runs it. The full procedure — tables, worked examples and the edge cases — is in the skill itself.
Where uncategorised harm actually shows up
Six places, roughly in order of yield.
What to look for specifically
Categories of emerging harm worth an explicit pass, because they recur across industries: An automation or AI agent behaving badly in a case nobody tested, A third-party failure you are carrying, A group struggling with something new.
Expect false positives, and say so
This scan trades precision for recall deliberately. A candidate list of fifteen where two are real is a success, and framing it that way protects the exercise — a scan presented as fifteen findings will discredit itself on the thirteen.
Cadence
Run it periodically rather than continuously — this is a review, not a monitor. Once a candidate is confirmed, it graduates: it gets a category, a monitor with a volume forecast, and an owner. The scan's output is new categories and new monitors, which is how the categorised system stops being last year's.
Related skills
Free and open source, and vendor-neutral — it reads the conversations from whichever helpdesk you already run. Browse all 149 skills · connect your helpdesk over MCP · source on GitHub
