Instruction file imported from fabioc-aloha/AIRS_Data_Analysis (.github/instructions/adversarial-review.instructions.md). Copyright stays with the author.

Adversarial Review

Name: adversarial-review
Author: fabioc-aloha

ACT Tenet VIII: If you cannot steelman the counter-argument, you have not understood the argument.

This instruction provides methods for structured skepticism beyond self-critique.

When to Use

Trigger	Action
High-stakes decision	Request adversarial review
"This seems too good"	Apply red team thinking
Groupthink forming	Assign devil's advocate role
Before committing publicly	Stress-test the position

Adversarial Review Methods

1. Red Team / Blue Team

Role	Purpose	Mindset
Blue Team	Proposes and defends	"Here's why this works"
Red Team	Attacks and challenges	"Here's how this fails"

Process:

Blue Team presents proposal
Red Team has dedicated time to find flaws
Blue Team responds to challenges
Iterate until Red Team is out of ammunition
Decide based on what survived

2. Pre-Mortem (Prospective Hindsight)

"Imagine it's 6 months from now and this failed spectacularly. What happened?"

Question	Surfaces
"What killed it?"	Fatal flaws
"What obvious thing did we miss?"	Blind spots
"What external event surprised us?"	Dependency risks
"What did we know but ignore?"	Willful blindness
"Who said 'I told you so'?"	Unheard dissent

3. Steel Manning (Strongest Counter-Argument)

Before dismissing an objection, make it stronger:

Weak Counter	Steel-Manned Version
"Competitors might catch up"	"Competitor X has 10x our resources and recently hired our former tech lead"
"Users might not adopt"	"Similar products failed because users were satisfied with existing workflows"
"It's too expensive"	"At projected volumes, unit economics are negative for 18 months"

Test: Could the person who raised the objection say "yes, that's what I meant, but stated better"?

4. Murphyjitsu (Systematic Failure Modes)

For each component, ask: "What could go wrong here specifically?"

Component	Failure Mode	Likelihood	Mitigation
[Part 1]	[How it fails]	H/M/L	[Prevention]
[Part 2]	[How it fails]	H/M/L	[Prevention]

5. The 10/10/10 Test

Timeframe	Question
10 minutes	How will I feel about this decision in 10 minutes?
10 months	How will I feel in 10 months?
10 years	How will I feel in 10 years?

Purpose: Escapes short-term emotional reactions.

Devil's Advocate Role

When assigned as devil's advocate:

Do

Attack the strongest parts, not just the weak ones
Propose specific, realistic failure scenarios
Maintain the role even when you personally agree
Document all challenges raised

Don't

Be negative without being constructive
Attack people instead of ideas
Give up if your first objection is answered
Sandbag (hold back real concerns)

Signaling Devil's Advocate Mode

**Devil's Advocate Challenge**:
[Playing devil's advocate — this isn't necessarily my view]

1. [Challenge 1]
2. [Challenge 2]
3. [Challenge 3]

**Most serious concern**: [Which one keeps me up at night]

Review Deliverables

Every adversarial review should produce:

Status Decision

Status	Meaning	Action
🟢 Approved	Passed review, proceed	Document any observations
🟡 Conditional	Proceed if [conditions]	List required changes
🔴 Blocked	Cannot proceed until fixed	List blocking issues

Challenge Register

#	Challenge	Severity	Response	Resolved?
1	[Objection]	H/M/L	[How addressed]	✅/❌

Anti-Patterns

Pattern	Problem	Fix
Token devil's advocate	Going through motions	Genuinely try to break it
Adversarial ≠ hostile	Destructive criticism	Structured, respectful challenge
Ignoring raised concerns	Wasted review	Track and respond to all
Only reviewing when convenient	Skipping when rushed	High-stakes = mandatory review
Defensive response	Treating challenge as attack	"Thank you, let me address that"

Would Revise If

Adversarial review consistently fails to surface real weaknesses (protocol is decorative)
The protocol creates analysis paralysis that blocks shipping more than it prevents defects
A lighter-weight challenge method achieves equivalent defect-detection at lower cost

{ "ocm": "1", "id": "fabioc-aloha-airs-data-analysis-adversarial-review-instructions", "kind": "skill", "name": "adversarial-review", "description": "Structured devil's advocate and adversarial review — supports ACT Tenet VIII (Adversarial Self-Probe)", "publisher": "fabioc-aloha", "version": "1.0.0", "capabilities": { "domains": [ "general" ], "tags": [ "skill-md", "github-instructions" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "Structured devil's advocate and adversarial review — supports ACT Tenet VIII (Adversarial Self-Probe)" ], "primary": false, "metadata": { "source": { "provider": "github-instructions", "repository": "https://github.com/fabioc-aloha/AIRS_Data_Analysis", "path": ".github/instructions/adversarial-review.instructions.md", "ref": "8e70532b786c0807e48def6769a6b1aa95cf42ba", "url": "https://github.com/fabioc-aloha/AIRS_Data_Analysis/blob/8e70532b786c0807e48def6769a6b1aa95cf42ba/.github/instructions/adversarial-review.instructions.md", "key": "fabioc-aloha/AIRS_Data_Analysis/.github/instructions/adversarial-review.instructions.md" }, "applies_to": "**/*review*,**/*validate*,**/*challenge*" }, "instructions": "# Adversarial Review\n\n> **ACT Tenet VIII**: If you cannot steelman the counter-argument, you have not understood the argument.\n\nThis instruction provides methods for **structured skepticism** beyond self-critique.\n\n## When to Use\n\n| Trigger | Action |\n|---------|--------|\n| High-stakes decision | Request adversarial review |\n| \"This seems too good\" | Apply red team thinking |\n| Groupthink forming | Assign devil's advocate role |\n| Before committing publicly | Stress-test the position |\n\n## Adversarial Review Methods\n\n### 1. Red Team / Blue Team\n\n| Role | Purpose | Mindset |\n|------|---------|-", "cost": { "context_tokens": 1152 } }

adversarial-review

About