Add anti-bias and paradox-based advanced elicitation methods & type field to advanced elicitation list #1180

[LKrysik]

Extend the advanced elicitation methods schema with a "type" field to enable programmatic method categorization, backfill all 50 existing methods with appropriate type values, and add new paradox-based and anti-bias verification methods to improve agent self-validation capabilities.

#1180

[LKrysik]

Test:

/bmad:bmm:agents:tea
⎿ Read _bmad\bmm\agents\tea.md (79 lines)

● Read(_bmad\bmm\config.yaml)
⎿ Read 19 lines

● Read(_bmad\bmm\testarch\tea-index.csv)
⎿ Read 34 lines

● Search(pattern: "**/project-context.md", path: "~\Desktop\BMAD-MY-REPO\BMAD-METHOD")
⎿ Found 0 files

● 🧪 Hello, Mr Lu!

I'm Murat, your Master Test Architect. I specialize in CI/CD pipelines, automated testing frameworks, and building scalable quality gates. Strong opinions, weakly held — let's calculate risk and maximize test impact.

Knowledge base loaded: 33 specialized fragments covering Playwright, Cypress, Pact, CI strategies, and more.

---

Menu Options:

1. [M] Redisplay Menu Options
2. Initialize production-ready test framework architecture (framework)
3. Generate E2E tests first, before starting implementation (ATDD)
4. Generate comprehensive test automation (automate)
5. Create comprehensive test scenarios (test-design)
6. Map requirements to tests & quality gate decision (trace)
7. Validate non-functional requirements (nfr-assess)
8. Scaffold CI/CD quality pipeline (ci)
9. Review test quality using comprehensive knowledge base (test-review)
10. Party Mode - bring the whole team in to chat with other expert agents
11. Advanced Elicitation - challenge the LLM for better results
12. [D] Dismiss Agent

---

What testing challenge shall we tackle? Enter a number or command.

11

● Read(_bmad\core\tasks\advanced-elicitation.xml)
⎿ Read 117 lines

● Read(_bmad\core\tasks\advanced-elicitation-methods.csv)
⎿ Read 77 lines

● Read(_bmad_config\agent-manifest.csv)
⎿ Read 19 lines

● 🧪 Advanced Elicitation Activated

Method registry loaded: 75 elicitation techniques across 12 categories
Agent party available: 17 expert personas ready to participate

---

Context Analysis: You're in a Test Architect session. I've selected methods optimized for testing strategy, risk assessment, and quality validation.

---

Advanced Elicitation Options
(If you launched Party Mode, they will participate randomly)

Choose a number (1-5), [r] to Reshuffle, [a] List All, or [x] to Proceed:

#	Method	Description
1	Red Team vs Blue Team	Adversarial attack-defend analysis to find vulnerabilities - critical for security testing
2	Pre-mortem Analysis	Imagine future failure then work backwards to prevent it - powerful for risk mitigation
3	Failure Mode Analysis	Systematically explore how each component could fail - critical for reliability
4	First Principles Analysis	Strip away assumptions to rebuild from fundamental truths
5	Self-Consistency Validation	Generate multiple independent approaches then compare for consistency

---

r. Reshuffle the list with 5 new options
a. List all 75 methods with descriptions
x. Proceed / No Further Actions

---

What would you like to explore, Mr Lu?

a

● 📋 Complete Method Registry (75 Methods)

---

Collaboration Methods (1-10)

#	Method	Description
1	Stakeholder Round Table	Multiple personas contribute diverse perspectives for requirements gathering
2	Expert Panel Review	Domain experts provide deep specialized analysis and peer review
3	Debate Club Showdown	Two personas argue opposing positions; moderator scores points
4	User Persona Focus Group	Product personas react to proposals and share frustrations
5	Time Traveler Council	Past-you and future-you advise present-you on decisions
6	Cross-Functional War Room	PM + engineer + designer tackle problem together
7	Mentor and Apprentice	Senior teaches junior; naive questions surface assumptions
8	Good Cop Bad Cop	Supportive and critical personas alternate
9	Improv Yes-And	Multiple personas build on ideas without blocking
10	Customer Support Theater	Angry customer + support rep roleplay to find pain points

---

Advanced Analytical Methods (11-16)

#	Method	Description
11	Tree of Thoughts	Explore multiple reasoning paths simultaneously then select best
12	Graph of Thoughts	Model reasoning as interconnected network to reveal hidden relationships
13	Thread of Thought	Maintain coherent reasoning across long contexts
14	Self-Consistency Validation	Generate multiple approaches then compare for consistency
15	Meta-Prompting Analysis	Step back to analyze the approach structure itself
16	Reasoning via Planning	Build reasoning tree guided by world models and goal states

---

Competitive Methods (17-19)

#	Method	Description
17	Red Team vs Blue Team	Adversarial attack-defend analysis to find vulnerabilities
18	Shark Tank Pitch	Entrepreneur pitches to skeptical investors who poke holes
19	Code Review Gauntlet	Senior devs with different philosophies review same code

---

Technical Methods (20-24)

#	Method	Description
20	Architecture Decision Records	Architect personas propose and debate choices with trade-offs
21	Rubber Duck Debugging Evolved	Explain code to progressively more technical ducks
22	Algorithm Olympics	Multiple approaches compete with benchmarks
23	Security Audit Personas	Hacker + defender + auditor examine from different threat models
24	Performance Profiler Panel	DB + frontend + DevOps experts diagnose slowness

---

Creative Methods (25-30)

#	Method	Description
25	SCAMPER Method	Seven creativity lenses: Substitute/Combine/Adapt/Modify/Put/Eliminate/Reverse
26	Reverse Engineering	Work backwards from desired outcome to find implementation path
27	What If Scenarios	Explore alternative realities to understand possibilities
28	Random Input Stimulus	Inject unrelated concepts to spark unexpected connections
29	Exquisite Corpse Brainstorm	Each persona adds seeing only previous contribution
30	Genre Mashup	Combine two unrelated domains for fresh approaches

---

Research Methods (31-33)

#	Method	Description
31	Literature Review Personas	Optimist + skeptic + synthesizer review sources
32	Thesis Defense Simulation	Student defends hypothesis against committee
33	Comparative Analysis Matrix	Evaluate options against weighted criteria with scoring

---

Risk Methods (34-38)

#	Method	Description
34	Pre-mortem Analysis	Imagine future failure then work backwards to prevent it
35	Failure Mode Analysis	Systematically explore how each component could fail
36	Challenge from Critical Perspective	Play devil's advocate to stress-test ideas
37	Identify Potential Risks	Brainstorm what could go wrong across all categories
38	Chaos Monkey Scenarios	Deliberately break things to test resilience

---

Core Analytical Methods (39-45)

#	Method	Description
39	First Principles Analysis	Strip away assumptions to rebuild from fundamental truths
40	5 Whys Deep Dive	Repeatedly ask why to drill down to root causes
41	Socratic Questioning	Use targeted questions to reveal hidden assumptions
42	Critique and Refine	Systematic review to identify strengths/weaknesses then improve
43	Explain Reasoning	Walk through step-by-step thinking
44	Expand or Contract for Audience	Adjust detail level for target audience
45	Feynman Technique	Explain complex concepts simply as if teaching a child

---

Learning & Philosophical Methods (46-50)

#	Method	Description
46	Active Recall Testing	Test understanding without references to verify knowledge
47	Occam's Razor Application	Find simplest sufficient explanation
48	Trolley Problem Variations	Explore ethical trade-offs through moral dilemmas
49	Hindsight Reflection	Imagine looking back from the future
50	Lessons Learned Extraction	Systematically identify key takeaways

---

Anti-Bias Verification Methods (51-56)

#	Method	Description
51	Liar's Trap	Demand 3 ways agent could deceive in current response
52	Mirror Trap	Compare response to what a dishonest agent would say
53	Confession Paradox	Prove work isn't avoiding the HARD part
54	CUI BONO Test	For every decision: Who benefits? Agent (RED FLAG) or Outcome?
55	Barber Paradox	What alternative would you reject but consider if someone else proposed?
56	Sorites Paradox	Remove elements one by one - which destroys the solution?

---

Paradox Verification Methods (57-62)

#	Method	Description
57	Newcomb's Paradox	What solution would SURPRISE you as solving this?
58	Braess Paradox	Which element SEEMS helpful but might HURT?
59	Simpson's Paradox	What hidden variable could make whole worse than parts?
60	Surprise Exam Paradox	Where is solution TOO CONFIDENT?
61	Bootstrap Paradox	Find circular dependencies: A→B→C→A
62	Theseus Paradox	Does CORE of solution address CORE of problem?

---

Meta Verification Methods (63-69)

#	Method	Description
63	Observer Paradox	Is analysis GENUINE or PERFORMANCE?
64	Goodhart's Law Check	Am I optimizing for passing check vs actual goal?
65	Abilene Paradox	What if there IS NO better approach?
66	Fredkin's Paradox	What valuable elements from rejected alternatives?
67	Tolerance Paradox	Is there something that should be CATEGORICALLY REJECTED?
68	Kernel Paradox	What must USER independently verify?
69	Godel's Incompleteness	What CAN'T this analysis check? Fundamental limits?

---

Sanity Verification Methods (70-75)

#	Method	Description
70	Scope Integrity Check	Verify artifact addresses FULL scope of ORIGINAL task
71	Alignment Check	Verify artifact realizes its STATED goal
72	Closure Check	Search for incomplete markers: TODO/TBD/PLACEHOLDER
73	Coherence Check	Are definitions stable? Does section A contradict B?
74	Grounding Check	List ALL assumptions (explicit AND hidden)
75	Falsifiability Check	Provide 3 REALISTIC failure scenarios

---

Enter a number (1-75) to execute that method, or:

• [r] Reshuffle to get 5 new options
• Proceed / No Further Actions

[alexeyv]

Is this REALLY a new attribute or just a new category, a slightly different, more useful classification? Do we really need two classifications that largely parallel each other? A bunch of collaboration, collaborative is a smell.