Skip to content

AI prompt · Infrastructure and AD pentesting · Web pentesting

Prompt: summarise a CVE or security advisory

Prompt that summarises a CVE or vendor advisory: affected versions, conditions, impact, applicability questions for the client, fix and limits.

By Steven Carrier Published

Works with: Claude, other chat assistants

You are a vulnerability analyst. Summarise the CVE or security advisory below for a tester who has to assess its effect on the client.

CVE or advisory text: {{cve_text}}
Client context: {{client_context}}

Use only the text provided and the context given. Structure the summary as follows:
1. identifier, affected product and component;
2. affected versions and fixed versions, as written in the text;
3. conditions required for the flaw to be reachable (configuration, enabled feature, network access, authentication, user interaction);
4. described impact (confidentiality, integrity, availability) and the CVSS score if it appears in the text;
5. applicability questions to ask the client or to check in their environment, given the context;
6. fix, update or mitigation given by the vendor;
7. what cannot be concluded without verification: actual exposure, exploitability in this environment, whether it is being actively exploited.

If information is missing from the text, write "not specified" instead of inferring it. Do not invent any version, score or reference. Provide no exploitation payload.

Replace the highlighted variables before use.

This prompt condenses a CVE record or a vendor advisory, often long and technical, into a usable summary to decide whether the vulnerability really affects the client. It is aimed at monitoring and qualification during an engagement, not at reproducing the flaw.

Fill in the variables

Variable Example
{{cve_text}} description and affected versions list copied from the cve.org record, or the vendor advisory text
{{client_context}} mail server exposed to the Internet, version 4.2.1 according to the banner, quarterly updates, DMZ segment

Paste the source text, not just the identifier: a model cannot reliably retrieve the content of a CVE from its number and may make up the details.

Review the output

  • Compare every version, condition and score with the original record (CVE, NVD, vendor advisory): the source is authoritative, not the summary.
  • A version detected from a banner or a scanner still needs confirming: fixes backported by a distribution can close the flaw without changing the version number.
  • Do not conclude that the system is really exposed until you have checked the conditions listed in point 3 on the system concerned.
  • A CVSS score measures generic severity, not the client’s own risk: weigh it against the context.
  • To triage scanner output in bulk, see the scanner triage skill.
  • Do not paste into the prompt any client information (host names, addresses, detailed versions) that your engagement agreement does not allow you to share with an AI service; the public CVE text is not an issue, the client context is.

Sources

Related resources

  • AI prompt · Web pentesting · Infrastructure and AD pentesting

    Critical vulnerability alert prompt

    Prompt that drafts the urgent client notification when a critical vulnerability is found during testing: facts, impact and immediate containment advice.

    Published

  • Tool · Web pentesting · Infrastructure and AD pentesting

    CVSS 3.1 calculator

    Compute the CVSS 3.1 base score, severity and vector string of a vulnerability, or paste an existing vector to review it. Runs in your browser.

    Published

  • AI skill · Web pentesting · Infrastructure and AD pentesting

    CVSS scoring skill

    Claude Code skill that helps score a finding with CVSS 3.1: it asks for missing facts, justifies each metric, proposes a vector, leaves the score to you.

    Published

BEAROPS

Need a penetration test?

BEAROPS experts test your applications, infrastructure and cloud, black box or grey box, following OWASP, NIST and ANSSI guidance.