Skip to content

AI prompt · Red teaming

Prompt: prepare the purple team debrief

Prompt that prepares a purple team debrief from the Red Team timeline and SOC alerts: detections, gaps, quick wins and ATT&CK techniques to verify.

By Steven Carrier Published

Works with: Claude, other chat assistants

You are a detection analyst running a purple team debrief between the Red Team and the SOC. Prepare the debrief material from the two sources below.

Exercise context: {{context}}
Red Team timeline (actions carried out, timestamped): {{red_team_timeline}}
Alerts and events seen by the SOC: {{soc_alerts}}
Available log sources: {{log_sources}}

Produce:
1. a table, one row per step of the timeline, with: timestamp, action carried out, candidate ATT&CK technique (name and identifier), status (detected, partially detected, not detected, undetermined), matching alert or event if any, delay between the action and the alert;
2. the list of detection gaps, each with the log source that could have covered it and the information missing to conclude;
3. quick wins: improvements achievable within a few days (rule to tune, logging to enable, threshold to review), ranked by estimated effort and expected benefit;
4. longer-term improvements, to be planned;
5. questions to ask the SOC during the debrief.

Rules:
- correlate only on what the timestamp and the event content show; if the link between an action and an alert is uncertain, classify the step as "undetermined" and say what would settle it;
- do not invent ATT&CK identifiers: if you are not sure of an identifier, write it in square brackets with the words "to be verified";
- a missing alert may come from missing logging as much as from a missing rule: distinguish the two hypotheses;
- stay factual and non-accusatory: the aim is to improve detection, not to judge the team.

Replace the highlighted variables before use.

This prompt prepares the debrief session of a Red Team exercise run in purple team mode: for each step carried out, did the SOC see it, if not why, and what can be fixed quickly. It structures the discussion without replacing the teams’ own log analysis.

When to use it

  • At the end of a Red Team exercise, once the timeline is consolidated and the SOC alerts have been exported.
  • To prepare a detection tuning workshop with the analysts, with a list of topics already sorted.
  • To track improvement from one exercise to the next: keep the table and compare the statuses.

Fill in the variables

Variable Example
{{context}} authorised three-week Red Team exercise on the corp.example.org domain, SOC informed at the end of the exercise
{{red_team_timeline}} D3 09:12 simulated attachment opened on a pilot workstation; D3 11:05 LDAP query of the administration groups; D5 14:30 access to a file share with a service account
{{soc_alerts}} D3 09:15 antivirus alert on the pilot workstation (medium severity); D5 14:50 "unusual sign-in" alert on the service account
{{log_sources}} domain authentication logs, EDR on workstations, file server logs, no detailed directory service logs

Provide timestamps in the same time zone and format for both sources. To map actions to techniques, use the MITRE ATT&CK mapping prompt; for how to organise the exercise, see preparing a Red Team exercise.

Review the output

  • Check every identifier and technique name on attack.mitre.org, and record the ATT&CK version used.
  • Cross-check each "not detected" status against the logs: the action may have been detected in a tool whose export was not supplied.
  • Check the calculated delays: a clock or time zone offset skews the gaps.
  • Have the quick wins validated by those who manage the rules: the model does not know your actual tooling.
  • Do not paste into the prompt any client information (host names, accounts, IP addresses, alert contents) that your agreement does not allow you to share with an AI service.

Sources

Related resources

  • Methodology · Infrastructure and AD pentesting · Red teaming

    Active Directory config review

    Defensive Active Directory review checklist for audits: privileged accounts, krbtgt, delegation, LAPS, legacy protocols, GPOs, ADCS and logging.

    Published

  • Cheatsheet · Infrastructure and AD pentesting · Red teaming

    Active Directory enumeration

    Authenticated AD enumeration for an internal test: domain, accounts, groups, password policy, shares, LDAP queries and BloodHound collection.

    Published

  • AI skill · Web pentesting · Infrastructure and AD pentesting · Red teaming

    Engagement scoping skill

    Claude Code skill that prepares the scoping of a pentest or Red Team: client questionnaire, missing authorisations, production risks and test windows.

    Published

BEAROPS

Ready for a red team exercise?

BEAROPS runs red team exercises lasting one to six months, built on MITRE ATT&CK, to put your detection and response to the test.