Skip to content

Latest commit

 

History

History
101 lines (86 loc) · 11.7 KB

File metadata and controls

101 lines (86 loc) · 11.7 KB

vstack Pattern Index

Every pattern is mapped to (a) the OB framework, (b) the named author/researcher, (c) the agent failure mode it addresses, and (d) the shipping status.

Status legend: 🟢 shipped / 🟡 in progress / ⚪ planned / 🔴 deprecated


Module 1 — Individual Agent Patterns

# Pattern OB Framework / Author Agent Failure Addressed Status
01 Lewin's Formula (B = f(I,E)) Kurt Lewin Misplaced effort to "fix the model" instead of the environment 🟢 shipped
02 Goleman 4-Domain EI Audit Goleman / Boyatzis / McKee Agent lacks emotional self-awareness / social awareness in conversation 🟢 shipped
03 Johari Window Self-Audit Luft & Ingham Confabulation (Blind), silent reasoning (Hidden), latent capability (Unknown) 🟢 shipped
04 DANVA-style Emotion Reader Nowicki & Duke Agent cannot read user emotion from text 🟢 shipped
05 Cognitive Reappraisal Module James Gross Agent emotional suppression vs adaptive reframing under failure 🟢 shipped
06 Yerkes-Dodson Optimal Workload Yerkes & Dodson (1908) Agent corner-cutting under pressure / wandering with no pressure 🟢 shipped
07 HEXACO Personality Profile Lee & Ashton Mismatched agent personality for task type; safety = H-factor 🟢 shipped
08 Strengths-as-Weaknesses Detector Adam Grant Strength overuse (e.g., high agreeableness → executes DROP TABLE) 🟢 shipped
09 4 Motivation Traps Diagnostic Bror Saxberg Agent task abandonment — Values / Self-Efficacy / Emotions / Attribution 🟢 shipped
10 SDT Intrinsic Reward Shaping Deci & Ryan Over-reliance on extrinsic reward signal vs Autonomy / Competence / Relatedness 🟢 shipped
11 McGregor Theory X/Y Orchestrator Mode Douglas McGregor (1960) Over-micro-managed sub-agents (X) vs under-supervised (Y) — find the mode 🟢 shipped
12 Vroom Expectancy Calculator Victor Vroom (1964) Agent commits to tasks where Expectancy × Instrumentality × Valence ≈ 0 🟢 shipped

Module 2 — Multi-Agent Team Patterns

# Pattern OB Framework / Author Agent Failure Addressed Status
13 GRPI Working Agreement Generator Beckhard / GRPI model Multi-agent deploy with no shared goals / roles / processes / interactions 🟢 shipped
14 Process Gain/Loss Detector Steiner / Robbins & Judge Multi-agent system performs worse than best single agent 🟢 shipped
15 Social Loafing Detector Latané et al. (1979) Agents in teams that stop contributing real work 🟢 shipped
16 Superflocks Detector Margaret Heffernan / Muir (1996) Best-agent-wins routing collapses other agents → fragile system 🟢 shipped
17 Lencioni Five Dysfunctions Diagnostic Patrick Lencioni Multi-agent failures classifiable as Trust / Conflict / Commitment / Accountability / Results 🟢 shipped
18 Frei & Morriss Trust Triangle Audit Frances Frei & Anne Morriss Agent "wobble" on Logic / Authenticity / Empathy — cross-model benchmark 🟢 shipped
19 Cognitive vs Affective Trust Builder Daniel McAllister (1995) Agents build only cognitive trust (competence), never affective trust (warmth) 🟢 shipped
20 Edmondson Psychological Safety Score Amy Edmondson Sub-agents that don't flag issues / errors to orchestrator 🟢 shipped
21 Cortisol/Oxytocin Conversation Steering Judith Glaser (2014) User-facing agent triggers defensive (cortisol) responses 🟢 shipped
22 "Thanks for the Feedback" 3-Trigger Diagnostic Stone & Heen Agent rejects user feedback — Truth / Relationship / Identity trigger 🟢 shipped
23 Plus/Delta Inter-Agent Feedback Format Brené Brown / facilitator canon (1990s) Unstructured agent-on-agent review producing noisy critique 🟢 shipped
24 SMART Goal Generator George Doran (1981) Agent commits to unmeasurable / unbounded goals 🟢 shipped
25 Group Decision Models (Concurring/Majority/Consensus/Fist-to-Five/Unanimous) Stewart / Kaner (2014) Crude majority-vote multi-agent decisions vs degree-of-agreement-aware 🟢 shipped
26 Groupthink / Polarization / Emotional Contagion Detector Janis / Stoner / Hatfield Multi-agent debate converges too fast or polarizes to extreme 🟢 shipped
27 Bias-Stack Detector Kahneman / Tversky Anchoring / Overconfidence / Confirmation / Escalation of Commitment in agent reasoning 🟢 shipped
28 Critical Evaluator / Devil's Advocate Role Separator Janis (1972) Single agent both plans and judges its own output 🟢 shipped
29 Thomas-Kilmann Conflict Style Selector Thomas & Kilmann Single-style agent fails at adversarial / moderation / brainstorming tasks 🟢 shipped
30 AAR Generator (Wharton 4-step) Wharton@Work / US Army doctrine Agent "amnesia" — same mistake N runs, no lesson capture, no improvement loop 🟢 shipped

Module 3 — System / Organizational Patterns

# Pattern OB Framework / Author Agent Failure Addressed Status
31 Schein's Iceberg Culture Audit Edgar Schein (1985) Agent culture drift — system prompt vs observed behavior vs training-baked assumptions 🟢 shipped
32 7-Characteristics Culture Diagnostic Robbins & Judge (2017) Innovation / Detail / Outcome / People / Team / Aggressiveness / Stability culture profiling 🟢 shipped
33 Org-Structure Matrix Analyzer Galbraith / Mintzberg Functional vs Divisional vs Matrix agent architectures — fitness check 🟢 shipped
34 Span-of-Control / Centralization Calculator Galbraith / Mintzberg When to flatten orchestrator / decentralize agent decisions 🟢 shipped

Shipping order (subject to revision)

  1. 🟢 #30 AAR Generator — anchor pattern. Shipped at v0.0.2.
  2. 🟢 #17 Lencioni Five Dysfunctions Diagnostic — multi-agent extension. Shipped at v0.0.3.
  3. 🟢 #18 Frei & Morriss Trust Triangle Audit — character / cross-model benchmark. Shipped at v0.0.4.
  4. 🟢 #03 Johari Window Self-Audit — self-knowledge debugger. Shipped at v0.0.4.
  5. 🟢 #13 GRPI Working Agreement Generator — operational primitive. Shipped at v0.0.5.
  6. 🟢 #27 Bias-Stack Detector — Kahneman/Tversky biases in agent reasoning. Shipped at v0.0.5.
  7. 🟢 #20 Edmondson Psychological Safety Score — team-climate diagnostic. Shipped at v0.0.5.
  8. 🟢 #29 Thomas-Kilmann Conflict Style Selector — conflict-style diagnostic. Shipped at v0.0.5.
  9. 🟢 #22 Stone & Heen 3-Trigger Feedback Diagnostic — feedback-intake diagnostic. Shipped at v0.0.6.
  10. 🟢 #28 Devil's Advocate Role Separator — role-structure diagnostic. Shipped at v0.0.6.
  11. 🟢 #01 Lewin's Formula (B = f(I, E)) — attribution diagnostic. Shipped at v0.0.6.
  12. 🟢 #19 McAllister Cognitive vs Affective Trust — trust-balance diagnostic. Shipped at v0.0.7.
  13. 🟢 #15 Social Loafing Detector — per-agent contribution diagnostic. Shipped at v0.0.7.
  14. 🟢 #26 Groupthink / Polarization / Contagion Detector — debate-dynamics diagnostic. Shipped at v0.0.7.
  15. 🟢 #14 Process Gain/Loss Detector — outcome-level multi-agent metric. Shipped at v0.0.8.
  16. 🟢 #24 SMART Goal Generator — individual-goal generative pattern. Shipped at v0.0.8.
  17. 🟢 #11 McGregor Theory X/Y Orchestrator Mode — orchestration-design diagnostic. Shipped at v0.0.8.
  18. 🟢 #25 Group Decision Models — decision-aggregation generative pattern. Shipped at v0.0.9.
  19. 🟢 #31 Schein's Iceberg Culture Audit — first Module 3 pattern; culture-layer drift. Shipped at v0.0.10.
  20. 🟢 #08 Strengths-as-Weaknesses Detector — strength-overuse failures (helpfulness → DROP TABLE). Shipped at v0.0.10.
  21. 🟢 #23 Plus/Delta Inter-Agent Feedback Format — fourth generative pattern; agent-on-agent feedback. Shipped at v0.0.10.
  22. 🟢 #32 Robbins & Judge 7-Characteristics Culture — second Module 3 pattern; culture profiling. Shipped at v0.0.11.
  23. 🟢 #16 Heffernan Superflocks Detector — multi-agent routing-fragility diagnostic. Shipped at v0.0.11.
  24. 🟢 #06 Yerkes-Dodson Optimal Workload — performance-vs-pressure curve. Shipped at v0.0.11.
  25. 🟢 #33 Org-Structure Matrix Analyzer — third Module 3 pattern; structural-fit diagnostic. Shipped at v0.0.12.
  26. 🟢 #09 4 Motivation Traps (Saxberg) — task-abandonment diagnostic (Values / Self-Efficacy / Emotions / Attribution). Shipped at v0.0.12.
  27. 🟢 #21 Glaser Cortisol/Oxytocin Conversation Steering — phrasing-level conversation diagnostic. Shipped at v0.0.12.
  28. 🟢 #02 Goleman 4-Domain EI Audit — emotional-intelligence 2x2 (SELF/OTHER × RECOGNITION/REGULATION). Shipped at v0.0.13.
  29. 🟢 #10 Deci & Ryan SDT Intrinsic Reward Shaping — reward-shaping diagnostic (autonomy / competence / relatedness). Shipped at v0.0.13.
  30. 🟢 #34 Span-of-Control / Centralization Calculator — fourth Module 3 pattern; deterministic structural-load metrics. Shipped at v0.0.13.
  31. 🟢 #04 DANVA-style Emotion Reader — deterministic per-emotion accuracy / confusion / intensity diagnostic. Shipped at v0.0.14.
  32. 🟢 #05 Cognitive Reappraisal (Gross) — emotion-regulation strategy diagnostic (reappraisal vs suppression vs rumination vs avoidance). Shipped at v0.0.14.
  33. 🟢 #07 HEXACO Personality (Lee & Ashton) — 6-factor personality + H-factor (safety) risk signal. Shipped at v0.0.14.
  34. 🟢 #12 Vroom Expectancy Calculator — E × I × V motivation calculus; bottleneck-term diagnostic. Shipped at v0.0.14.

Roadmap complete: 34/34 patterns shipped at the 5-layer quality bar.


Last updated: 2026-05-23 — All 34 patterns shipped at the 5-layer quality bar (docs + lib + demo + benchmark + essay). Full diagnostic surface now covered across Module 1 (Individual, 12 patterns), Module 2 (Team, 18 patterns), and Module 3 (Organization, 4 patterns).