escalation_skill
- Python
- Official
7.4k
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill anthropics/knowledge-work-plugins --skill escalation- SKILL.md8.1 KB
Overview
This skill structures and packages support escalations so engineering, product, security, or leadership can act immediately. It creates concise briefs with business impact, exact reproduction steps, mitigation attempts, and a clear ask. Use it to decide whether to escalate, draft the escalation, and manage follow-up until resolution.
How this skill works
The skill inspects the incident context, support history, and impact dimensions to recommend escalation tier and severity. It produces a standardized escalation brief containing impact metrics, a 3–5 sentence issue summary, step-by-step reproduction, evidence, actions tried, and the specific decision or fix requested. It also recommends follow-up cadence and de-escalation conditions.
When to use it
- Issue appears to need code changes, infrastructure investigation, or security review
- Multiple customers or high-value accounts are impacted or SLA risk exists
- Support cannot reproduce or lacks required access/privileges to diagnose
- Customer requests executive involvement or legal/PR risk is present
- Issue has been open beyond SLA or is recurring/patterned
Best practices
- Quantify impact: customers affected, revenue at risk, SLA status
- Provide exact reproduction steps with environment, inputs, and evidence
- Be explicit in the ask: investigate, fix, approve, or prioritize — include a deadline
- Log what’s been tried and ruled out to avoid duplicated work
- Assign follow-up cadence by severity and own the customer communication
- Include links to tickets, logs, and internal threads for quick context
Example use cases
- Frontline support confirms a bug that blocks exports for multiple customers and needs engineering triage
- Senior support identifies a recurring workflow gap and escalates to Product with frequency and use cases
- Suspected data exposure is reported and must be escalated directly to Security with containment details
- High-revenue customer threatens churn after a missed SLA; escalate to Leadership with impact and decision needed
- Support needs a rapid hotfix request: include reproduction, logs, and a same-day deadline
FAQ
Precise reproduction steps from a clean state, including environment, exact inputs, and evidence — this saves engineering time.
How do I choose severity?
Map severity to breadth, depth, duration, and revenue/reputation risk: Critical for production down or data risk; High for major functionality broken or SLA at risk; Medium when a workaround exists.