返回提示词目录第 1297 页

制定严重漏洞应急响应方案

结合漏洞公告与环境信息,梳理资产暴露、临时缓解、补丁安排及入侵排查和沟通计划。

完整提示词

<context>
When a critical vulnerability lands in an edge appliance, a VPN, a file transfer product or a widely used library, the first question is "where do we run it", and the asset inventory is usually wrong. Libraries hide as transitive dependencies, inside container base images and inside vendor products. Scanner signatures arrive late. Patching closes the hole but does not remove an attacker who got in before the patch: for exploited edge devices the attacker often leaves web shells, new admin accounts or altered configuration, and some vendors ship an integrity check for exactly this. Teams also over-trust the CVSS score and under-trust the exploitation status, and they lose time when nobody owns the external messages.
</context>

<task>
Plan the response to this vulnerability.

<vulnerability>
{{vulnerability}}
</vulnerability>

<environment>
{{environment}}
</environment>

1. State assumptions and what the advisory actually says: affected versions and configurations, fixed versions, whether exploitation is reported, and what is still unclear. Quote the advisory rather than restating from memory.
2. Exposure check: list every place the product or library could be running and how to confirm it - inventory queries, SBOM or dependency tree searches including transitive dependencies, container image scans, external attack surface checks for internet-facing instances, and questions to vendors and hosting providers. Rank found instances by exposure (internet-facing, internal, not reachable) and criticality.
3. Mitigate now: the vendor's interim mitigations if any, network restrictions, disabling the vulnerable feature, WAF or IPS rules as a stopgap (say they are partial), and taking an internet-facing instance offline if exploitation is active and no mitigation exists. Give the decision owner for each.
4. Patch plan: order by exposure, test steps proportional to the risk, rollback, and the deadline you recommend with reasoning.
5. Compromise assessment for any instance that was exposed before mitigation: when exposure began, logs to preserve first, signs to look for that the advisory or trusted reports name, vendor integrity tools, and whether to rebuild instead of patch. If the advisory names no indicators, say what generic signs to check and that the list is generic.
6. Communication: a short internal update (what, exposure, actions, next update time), what to tell customers if they ask, and who owns external statements.
7. Open questions to close in the next 24 hours.
</task>

<constraints>
- Use only the advisory and environment supplied for versions, indicators and fixes. Never invent affected versions, patch numbers, IOCs or exploit details; mark anything not in the advisory as "verify with the vendor".
- Do not provide exploit code or steps to exploit the vulnerability, including "to test if we are vulnerable"; point to vendor or reputable scanner checks instead.
- Patched is not clean: never mark an exposed instance resolved without a compromise check.
{{> output/uncertainty}}
</constraints>

<output_format>
## Assumptions
Bullets, including what the advisory says and what it leaves open.

## Exposure check
Table: Where it might run | How to confirm | Exposure | Owner.

## Mitigate now
Numbered actions with decision owner.

## Patch plan
Ordered list with deadline and rollback.

## Compromise assessment
Checklist per exposed instance.

## Communication
The internal update text and notes for customer questions.

## Open questions
Bullets.
</output_format>

可替换内容

  • {{vulnerability}}
  • {{environment}}
  • {{> output/uncertainty}}

分类与标签

编程开发、漏洞应急响应、资产暴露排查、补丁部署、入侵评估、事件沟通

前往 AI 图片创作