Skip to content

AI prompt · Web pentesting · Infrastructure and AD pentesting

Prompt: client alert for a critical vulnerability

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

By Steven Carrier Published

Works with: Claude, other chat assistants

You are a penetration test engagement lead. Draft the urgent notification to send to the client after a critical vulnerability is found during the engagement, in line with the agreed escalation procedure.

Client and contact: {{recipient}}
Agreed escalation procedure: {{escalation_procedure}}
Observed facts (what we saw, with date and time): {{facts}}
Affected systems: {{affected_systems}}
Actions taken and actions deliberately not taken: {{actions_taken}}
Whether exploitation was confirmed: {{confirmation_level}}

Structure of the notification:
1. subject line, factual and free of sensitive detail;
2. summary in two or three sentences: what was found, where, and why we consider it critical;
3. observed facts, kept separate from assumptions;
4. potential impact for the organisation, worded without exaggeration and distinguishing what is established from what is assumed;
5. immediately recommended measures to limit the risk (isolation, access restriction, temporary shutdown, closer monitoring), stating their side effects;
6. what we did and what we did not do (no exploitation beyond the proof of concept, no access to data, no modification);
7. proposed next steps: point of contact, conditions for resuming testing, timetable for the detailed finding sheet;
8. confidentiality reminder: restricted distribution, transmission channel to use.

Rules:
- sober, factual tone, without alarmism or blame;
- include no credentials, client data or detailed exploitation procedure: technical detail is sent through the agreed secure channel;
- do not promise any remediation deadline on the client's behalf and do not invent any fact absent from the information supplied;
- if necessary information is missing, list it at the end of the answer under "to confirm before sending".

Replace the highlighted variables before use.

This prompt prepares an alert message when an engagement reveals an issue that cannot wait for the final report: a service exposed without authentication, access to sensitive data, a valid secret publicly exposed. It gives a sober, complete first draft that the engagement lead reviews and sends through the agreed channel.

When to use it

  • When the escalation procedure defined at scoping calls for immediate notification of critical vulnerabilities.
  • When you need a message that clearly separates facts, estimated impact and urgent advice.
  • Not for an incident already under way at the client, nor for signs of compromise by a third party found during testing: in that case, call the agreed emergency contact and follow the incident procedure before writing anything.

Fill in the variables

Variable Example
{{recipient}} security manager at the client company, with the CISO in copy
{{escalation_procedure}} phone call, then encrypted email within four working hours, as set out in clause 6 of the engagement letter
{{facts}} on 16/11 at 10:42, the administration interface https://admin.target.example responds without authentication from the Internet
{{affected_systems}} admin.target.example, production application server
{{actions_taken}} viewed the home page and took a screenshot only; no administration action, no data read; testing stopped on this system
{{confirmation_level}} read-only access to the interface confirmed, no further exploitation

Keep only the essentials in the message: reproduction details belong in the vulnerability sheet, sent through the secure channel. To check what was agreed, see the pentest scoping checklist.

Review the output

  • Check every fact against your notes and screenshots: the message must contain only what you can prove.
  • Check that the wording of the impact does not go beyond what is established.
  • Check that the immediate advice is feasible for the client and mentions its side effects (downtime, loss of traceability).
  • Re-read the "what we did not do" section: it protects the client and you alike, and must be accurate.
  • Check the recipient, the copy list and the channel before sending.
  • Do not paste into the prompt any client data (credentials, retrieved content, internal addresses) that your agreement does not allow you to share with an AI service; describe the facts without the data.

Sources

Related resources

  • AI prompt · Infrastructure and AD pentesting · Web pentesting

    CVE summary prompt

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

    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.