A strategy, plan, architecture or solution exists in reviewable form so the Red Team can attack concretely.
Red Teaming
Prerequisite
What needs to be finished first
First risk hypotheses are available so the Red Team can start from real weak points.
Preparation
What needs to be ready before start
Complete documents for the artifact (strategy, architecture, plan); whiteboard or Miro with attack table (attack vector, effect, probability, mitigation); STRIDE/PASTA templates for security if needed.
One Red Team (3-5 people), ideally external to the original team; one original-team representative for understanding questions (not defense); one moderator; one scribe.
Provide artifact 2-3 days before workshop; clarify context and assumptions; known attack patterns in domain (security: OWASP; business: competitor moves); time limit for Red Team preparation.
1-2 days (1 day preparation, 4-8 h workshop)
Red Team works autonomously on attack strategies before workshop. Workshop has clear rules: original team does not defend, but listens and notes. Q&A only after complete presentation.
Core question
The one question this method answers
Which attack vectors, weaknesses or entry paths overcome the artifact, and which of these vectors are likely enough to justify mitigation?
Flow
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Red Team preparation | 4-8 h | Red Team studies the artifact. Generates attack tree or list. Per attack: vector, required capabilities, effect, probability, existing mitigation. | Red Team needs time and distance. Without preparation, Red Teaming becomes discussion round. External people are often more creative than internal ones. |
2Phase 2: Context and rules | 30 min | Moderator briefly presents artifact. Red Team and original team ask understanding questions. Rules clear: original team listens, answers facts, does not defend. | Defense reflex is the most common method error. Clear rule: discussion only in Phase 4. |
3Phase 3: Attack presentation | 2-3 h | Red Team presents attacks systematically. Per attack: vector, scenario, effect. Original team asks only for understanding, not rebuttal. | If original team interrupts and defends, moderator interrupts back. Presentation must be complete before evaluation starts. |
4Phase 4: Evaluation and prioritization | 90 min | Per attack, evaluate effect, probability and mitigation effort. Original team adds context that changes probability. Top attacks go to risk register as risks. | Probability is subjective. Consensus is not required, but rationale must be documented. |
5Phase 5: Mitigations and follow-ups | 60-90 min | Sketch mitigation per top attack (prevention, detection, response). Owner and deadline. Log changes to artifact. | Mitigation must be concrete. "More monitoring" without detail is not mitigation. Owner mandatory. |
Artifact
What comes out at the end
Red Team report with artifact description, attack list (vector, scenario, effect, probability, mitigation effort), prioritized risks and mitigations with owner and deadline. Supplemented by methodology notes and open questions.
One entry per Red Team session with date and artifact version. Mitigations tracked separately over time. New Red Team run for major changes to artifact.
- Confluence or Notion page with risk register
- Markdown in repo under docs/security/redteam-reports/
- Spreadsheet (Sheets, Excel) for attack table
- Threat-modeling tools such as ThreatDragon or Microsoft Threat Modeling Tool for security red teaming
- Miro board with attack tree
red-teaming-working-template.md
Compact working template for Red Teaming with challenge scope, attack points, findings, and mitigations.
Red Teaming Working Template
Goal
Challenge assumptions, expose weak spots, and strengthen the plan, strategy, or design.
Context
When and for what do we use this method?
Input
Which data, observations, decisions, or materials are available?
Working area
- Scope:
- Attack points:
- Weak assumptions:
- Findings:
- Mitigations:
Output artifacts
- Challenge notes:
- Risk register:
- Mitigation plan:
Open questions
- ...
Next step
Owner, date, and success signal.
Example output
Concrete filled scenario, fictional example
red-teaming-beispiel.md
Concrete filled scenario, fictional example
Red Team Report - Workspot Booking API v2 Architecture, May 2026
Red Team: 3 external architects + 1 internal security champion. Preparation 6 h.
Top attack vectors
A1: Partner load spike through leaked API key
- Scenario: partner API keys land in public GitHub repo. Bot traffic generates 50,000 requests/min.
- Effect: Booking API overloaded, p95 breached, extra cost approx. 8,000 EUR/day, reputation damage.
- Probability: High (several partners are solo self-employed without strict secret hygiene).
- Mitigation: per-partner rate limit (Owner @ben, by 2026-06-30), token rotation API (Owner @anna, by 2026-07-15), public repo scanner for secrets (Owner @marcus, by 2026-07-30).
A2: Inventory race in multi-region booking
- Scenario: two bookings for same workspace confirmed simultaneously in different regions.
- Effect: double booking, customer dispute, manual rollback.
- Probability: Medium (only arises with active-active multi-region).
- Mitigation: sticky region per workspace ID instead of tenant (Owner @lisa, adapt ADR-014).
A3: Notification worker as single point of failure
- Scenario: Kafka consumer hangs, notifications accumulate, confirmation emails delayed.
- Effect: activation rate drops, support volume rises.
- Probability: Medium.
- Mitigation: consumer in 2 replicas + alert at lag >100 (Owner @marcus, by 2026-06-22).
Further attacks (excerpt, without immediate mitigation): cache invalidation bug, compliance audit findings in eu-west region, auth service downtime.
Follow-up: Top 3 into risk register Booking API, ADR-014 adjusted, reviews in 30 days.
Pitfalls
Recognize symptoms and steer against them
Original team defends immediately
Presentation becomes discussion, Red Team cannot get through attack list.
Moderator interrupts. Clear rule: understanding questions yes, defense no. Evaluation only in Phase 4.
Red Team knows artifact too little
Attacks are generic, do not hit artifact.
Preparation time mandatory. Red Team must have artifact 2-3 days beforehand. If time window is short, postpone.
Internal Red Teaming
Red Team and original team share assumptions, blind spots remain.
Bring external people into Red Team, ideally with different perspective. If impossible, at least another department.
Attacks without mitigation
Long attack list, nobody takes mitigation ownership, list gathers dust.
Top attacks require mitigation, owner, deadline. Other attacks as watchlist with next review date.
Politics instead of method
Red Team is used to fight internal conflicts, attacks are hidden career moves.
Independent moderator. Attacks relate to artifact, not people. Moderator interrupts personalization.
Stop criteria
Done signals checkable in under a minute
Finished the runsheet?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.