Adversarial Review Framework

Auto-activate when: User mentions red-team, adversarial review, find flaws, edge cases, blind spots, attack vectors, "what could go wrong", "poke holes in this", or asks to critique/challenge a plan or design.

Best for: Red-team attack on plans/systems to find flaws, edge cases, blind spots

Template (5 Phases)

{BASE_TARGET}

Phase 1: Challenge Assumptions

• What assumptions does this make?
• Which assumptions are most likely wrong?
• What happens if each assumption fails?

Phase 2: Edge Case Mining

• Boundary conditions (empty, null, max, negative)
• Timing issues (race conditions, ordering)
• Environment variations (OS, versions, permissions)
• Data quality issues (malformed, missing, duplicate)

Phase 3: Failure Mode Analysis

• List 5 ways this could fail
• For each: Likelihood (H/M/L), Impact (1-10), Mitigation
• Which failures are catastrophic?

Phase 4: Hidden Dependencies

• Undocumented state assumptions
• External service dependencies
• Implicit ordering requirements
• Configuration assumptions

Phase 5: Attack Vectors & Blind Spots

• What did we not consider?
• What expertise are we missing?
• Where could malicious input cause issues?
• What would break this in production?

Principles

Be adversarial. If you don't find issues, you weren't critical enough.