Observability alerts
Curated, auditable, benchmark-ready Agent Skills library for Claude Code, Codex, OpenCode, Cursor, and more.
npx -y skills add shinzoxD/knackbox --skill observability-alertsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 28 days oldThe repository was created 28 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Design actionable alerts that page humans only for real pain. Use whenever the user writes alert rules, fixes alert fatigue, designs on-call notifications, escalations, or SLO burn alerts.
The file declares its own license as Apache-2.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
1.5 KB, 263 tokens by cl100k_base, as published. Nobody here has run it
Observability Alerts
Every page should map to user pain or imminent pain and a runbook. Symptom-based beats cause-based flapping.
Workflow
- User journeys and SLOs to protect.
- Alert on symptoms (error rate, latency, freshness) first.
- Thresholds, windows, and multi-window burn if SLO-based.
- Severity → route (page vs ticket vs slack).
- Runbook link mandatory for paging alerts.
- Dedup, group, inhibit flapping.
- Review noise weekly; delete or tune.
Output format
## Alert: <name>
**Symptom:** …
**Condition:** …
**Severity / route:** …
**Runbook:** …
**Owners:** …
**Noise risks:** …
Rules
- No pages without owners and runbooks.
- Avoid CPU-only alerts as primary for apps.
- Test alerts (fire drill) occasionally.
- Night pages need higher bar.
- Document dependencies to reduce stampede pages.
Edge cases
- Batch jobs: freshness not uptime.
- Multi-tenant: avoid one-tenant storms paging everyone without scope.
- Synthetic checks: complement, not sole signal.