返回提示词目录第 1400 页

编写组织漏洞管理制度

结合组织资产与人员能力,起草涵盖漏洞分级、修复期限、例外审批和实施安排的制度。

完整提示词

<context>
Vulnerability management policies fail in two ways. Some are copied templates that promise deadlines the organisation has never met, so every audit finds non-compliance and every team ignores the policy. Others rank everything by CVSS base score alone, which floods teams with "critical" findings that are not reachable while an exploited medium on an internet-facing system waits. A workable policy defines scope with real asset classes, requires authenticated scanning or agent coverage and measures it, prioritises with exploitation evidence (a known-exploited list such as CISA KEV, an exploit prediction score such as EPSS), exposure and asset criticality on top of severity, sets deadlines the teams have agreed they can meet, and treats exceptions as time-boxed risk acceptances signed by someone with authority over the risk. It also names who owns each asset class, because "IT will patch it" is not an owner.
</context>

<task>
Write a vulnerability management policy for:

<organisation>
{{organisation}}
</organisation>
{{#existing_practice}}

<existing_practice>
{{existing_practice}}
</existing_practice>
{{/existing_practice}}
{{#frameworks}}Frameworks or obligations: {{frameworks}}{{/frameworks}}

1. Decisions you need to make: list the choices the organisation must confirm before adopting the policy (deadlines per priority, who can accept risk at each level, scan frequency, scope exclusions), with a recommended default and the trade-off for each.
2. Draft policy with these parts:
   - purpose and scope, by asset class, including third-party and cloud services where the organisation controls configuration;
   - roles: policy owner, asset owners, remediation owners, security team, risk approvers;
   - discovery: asset inventory as the base, scanning and agent coverage targets, dependency and container image scanning for own software, frequency, and a coverage metric;
   - prioritisation: a small matrix combining severity, exploitation evidence, exposure and asset criticality into three or four priority levels;
   - remediation deadlines per priority level, from the decisions above, labelled as defaults to confirm, plus an emergency process for actively exploited flaws;
   - exceptions and risk acceptance: when allowed, required compensating controls, approver by level, maximum duration and renewal;
   - verification: rescanning to close a finding;
   - metrics and reporting: coverage, deadline compliance, mean time to remediate by priority, open exceptions, aging;
   - review cycle and policy ownership.
3. If frameworks are named, add a short mapping of policy sections to the relevant requirements, with every clause reference marked `[VERIFY]`.
4. Rollout notes: how to phase it in (measure first, then enforce), what to communicate to teams, and how to handle the existing backlog without breaching the new policy on day one.
</task>

<constraints>
- Write deadlines as proposals for the organisation to confirm; do not present them as an industry standard.
- Never invent framework clause numbers or quote requirements as fact; mark them `[VERIFY]`.
- Keep the policy to what the organisation can run with its stated tools and people; flag any requirement that needs a capability they do not have.
- Plain language. No filler about the importance of security.
{{> output/uncertainty}}
</constraints>

<output_format>
## Decisions you need to make
Table: Decision | Recommended default | Trade-off.

## Draft policy
The policy as numbered sections with short paragraphs, the prioritisation matrix as a table and the deadlines as a table.

## Rollout notes
Numbered steps.
</output_format>

可替换内容

  • {{organisation}}
  • {{#existing_practice}}
  • {{existing_practice}}
  • {{/existing_practice}}
  • {{#frameworks}}
  • {{frameworks}}
  • {{/frameworks}}
  • {{> output/uncertainty}}
  • [VERIFY]

分类与标签

办公效率、漏洞管理、安全制度、风险审批、修复流程

前往 AI 图片创作