Organize a security investigation

Use this to organize sanitized observations when you need a disciplined next step. It separates what the logs show from explanations that still need evidence.

Prompt template

Using the sanitized observations below, propose a triage plan. Separate observed facts, possible explanations, and missing evidence. Prioritize reversible, read-only checks. Do not invent indicators or attribute an attacker. State when escalation to an incident responder is appropriate. Observations: [observations]

Boundary: Produce a triage plan only. Keep facts, names, numbers, timestamps, and meaning intact; leave observations, systems, and settings unchanged. List proposed checks and changes separately. Don't execute commands or take containment actions.

Fill in this prompt in the builder

What to supply

  • Sanitized observations, relevant timestamps, and what is actually known.
  • Any approved evidence sources and operational limits inside the observation text.

Worked example

Illustrative example: the input and result below are invented.

Input

Eli, 28, has a sanitized SIEM summary: six failed sign-ins for a lab account between 09:10 and 09:12 UTC, then one successful sign-in at 09:13. The source address is omitted. MFA and session details aren't included. No compromise is confirmed.

Filled-in prompt

Show the complete prompt
Using the sanitized observations below, propose a triage plan. Separate observed facts, possible explanations, and missing evidence. Prioritize reversible, read-only checks. Do not invent indicators or attribute an attacker. State when escalation to an incident responder is appropriate. Observations: Lab account: six failed sign-ins from 09:10 to 09:12 UTC, then one success at 09:13 UTC. Source address omitted. MFA and session details unavailable. No compromise confirmed. Approved approach: read-only evidence review; no account or policy changes.

Boundary: Produce a triage plan only. Keep facts, names, numbers, timestamps, and meaning intact; leave observations, systems, and settings unchanged. List proposed checks and changes separately. Don't execute commands or take containment actions.

Result

Observed: The supplied summary contains failed sign-ins followed by a success.
Possible explanations: A user corrected a password, an automated process retried, or unauthorized access occurred. None is established by this sequence alone.
Read-only checks: Retrieve the relevant authentication records; inspect MFA outcome, source context, session identifiers, and activity after the successful sign-in. Confirm timestamp alignment.
Missing evidence: Source context, MFA detail, and subsequent activity.
Escalation: Follow the incident-response process if the reviewed evidence indicates unauthorized access or consequential activity.
Changes: None; observations retained as supplied.

What I’d check or correct

I'd check that the assistant didn't call this an attack based only on the failed-then-successful sequence. Any proposed account disablement would exceed the supplied read-only scope.

What to check

  • Observed facts, possible explanations, and missing evidence are separate.
  • The plan preserves the observations, names, numbers, and meaning; it lists proposals without changing settings or taking containment actions.
  • Read-only checks are prioritized, with escalation tied to evidence and the established response process.

Related pages

Share this page