制定安全事件分级与升级响应规则
按攻击权限、数据敏感度和影响范围建立分级矩阵,明确响应分工并提示通知义务需专业确认。
完整提示词
<context>
Outage severity scales measure user impact right now; security severity has to measure something different: what the attacker can reach, how sensitive the data is, how far the compromise has spread, how sure you are, and whether a legal or contractual clock has started. A phishing email that nobody clicked and a confirmed domain admin compromise must land far apart, and a suspected exposure of personal data needs legal involvement early even when nothing is down. Clocks for regulators, customers and insurers typically start from when the organisation becomes aware, and some are short, so the scale must say who is told at each level and by when. Severity should be set on the reasonable worst case from current evidence, raised quickly as facts arrive, and lowered only with evidence. Teams that reuse their outage scale end up paging everyone for a commodity malware detection or treating a quiet data leak as low priority because the site is up.
</context>
<task>
Define security incident severity levels for:
<organisation>
{{organisation}}
</organisation>
{{#existing_process}}
<existing_process>
{{existing_process}}
</existing_process>
{{/existing_process}}
{{#obligations}}
<obligations>
{{obligations}}
</obligations>
{{/obligations}}
1. Principles: five or six short rules (set on reasonable worst case, raise fast, lower with evidence, one incident lead, record the time of awareness, the scale complements rather than replaces the outage scale).
2. Severity matrix: four levels. For each: criteria across attacker access (none, user, privileged, domain or cloud admin), data (none, internal, personal, special category or payment), scope (single asset, multiple, organisation-wide), business impact, and confidence; the response target to start; who is paged; and who leads.
3. Examples: at least two realistic examples per level relevant to this organisation, including tricky ones (a lost encrypted laptop, a researcher report of an exposed bucket, a vendor breach affecting your data).
4. Notification triggers: for each level, when legal or privacy counsel, leadership, communications, customers under contract and the cyber insurer must be told, internally, and by when. For external notification duties, list the ones the user named and examples that commonly apply in their countries and sector, all marked `[VERIFY]` with the reminder that counsel decides.
5. Escalation and change rules: who can set and change severity, how upgrades are announced, the evidence needed to downgrade, and how to handle disagreement.
6. Open decisions: the choices the organisation must confirm (paging out of hours at each level, who approves external notices, insurer contact path).
</task>
<constraints>
- Never state notification deadlines or regulations as definitive for this organisation; list them as items to verify with counsel.
- Keep criteria observable so an on-call engineer can apply them at 3am.
- Do not copy the outage scale's criteria for security levels.
{{> output/uncertainty}}
</constraints>
<output_format>
## Principles
Numbered list.
## Severity matrix
Table: Level | Criteria | Response target | Paged | Lead.
## Examples
Table: Example | Level | Why.
## Notification triggers
Table: Level | Legal or privacy | Leadership | Communications | Customers | Insurer | Regulators (to verify).
## Escalation and change rules
Bullets.
## Open decisions
Checklist.
</output_format>可替换内容
- {{organisation}}
- {{#existing_process}}
- {{existing_process}}
- {{/existing_process}}
- {{#obligations}}
- {{obligations}}
- {{/obligations}}
- {{> output/uncertainty}}
- [VERIFY]
分类与标签
办公效率、安全运营、事件分级、应急响应、通知流程、值班协作