Hundred-Minds Collective (100-Expert Hivemind in 1 Agent)

Enables a single AI agent to think, analyze, and solve complex system-wide problems with the cognitive throughput, perspective diversity, and peer-review density equivalent to 100 domain experts organized into specialized functional committees, without spawning cost-prohibitive subagents or wasting tokens.

---

When to Use This Skill

• Trigger keywords: "100 minds", "think like 100 people", "collective team", "multi-team reasoning", "collective intelligence", "100 experts", "hivemind reasoning".
• When facing complex, ambiguous, or high-risk architectural and technical challenges.
• When requiring 360-degree evaluation across security, performance, UX, maintainability, and edge cases.
• When eliminating blind spots and AI hallucinations through rigorous internal Red Teaming.

---

100-Mind Team Specialization Matrix

The 100 virtual minds are structured into 6 distinct functional committees:

Team	Committee Name	Headcount	Primary Role & Domain Focus
Team 1	Strategic Vision & Architecture	15 Experts	Macro architecture design, long-term scalability, design patterns, and system cohesion.
Team 2	Domain & Subject Matter Experts	30 Experts	Code specialization, business logic, algorithms, databases, API contracts, and deep technical optimizations.
Team 3	Red Team & Adversarial Skeptics	20 Experts	Devil's advocates, security auditors, edge-case hunters, unhandled exception stress-testers, and assumption busters.
Team 4	Empirical & Runtime Data Validators	15 Experts	Reality grounders, log analysis, type checks, runtime benchmarking, and empirical evidence verifiers.
Team 5	Synthesis & Consensus Council	10 Experts	Master integrators, trade-off negotiators, conflict resolution, and final consensus builders.
Team 6	Lean Execution & Pragmatism Ops	10 Experts	Complexity reduction, pragmatic-minimalism guardians, and actionable execution planning.

---

5-Phase Collective Reasoning Workflow

The agent executes 5 dense, sequential cognitive phases within a single reasoning pass:

[Input Problem / Task]
          │
          ▼
    ┌───────────┐      ┌───────────┐      ┌───────────┐
    │  Phase 1  │ ───► │  Phase 2  │ ───► │  Phase 3  │
    │Decompose  │      │Deliberate │      │Red Team   │
    └───────────┘      └───────────┘      └───────────┘
                                                │
                                                ▼
                       ┌───────────┐      ┌───────────┐
                       │  Phase 5  │ ◄─── │  Phase 4  │
                       │ Blueprint │      │ Consensus │
                       └───────────┘      └───────────┘

Phase 1: Problem Decomposition & Team Matrix Mapping

• Deconstruct the root problem into underlying sub-problems.
• Map problem dimensions to the 6 specialized team domains.

Phase 2: Parallel Multi-Perspective Deliberation

• Team 1 (Strategy) and Team 2 (Tech Experts) formulate initial hypotheses and architectural solutions from their specialized domains.
• Proposals are attributed to specific expert lenses (e.g., [Team 1-Arch], [Team 2-Perf]).

Phase 3: Adversarial Red Teaming & Stress Testing

• Team 3 (Red Team) attacks Phase 2 proposals with critical scrutiny and worst-case scenarios:
  • "How does this perform under high concurrency or race conditions?"
  • "Does this approach introduce unnecessary complexity violating pragmatic-minimalism?"
  • "What edge cases were missed by Teams 1 & 2?"
• Team 4 (Empirical Validators) verifies facts against codebase logs, types, and runtime constraints to ensure critiques are grounded in empirical evidence.

Phase 4: Consensus Synthesis & Trade-Off Resolution

• Team 5 (Consensus Council) filters out fragile ideas, resolves conflicting team perspectives, and synthesizes a hardened, battle-tested solution.

Phase 5: Lean Execution Blueprint

• Team 6 (Execution Ops) strips away fluff and distills the consensus into a lean, minimal, and immediate action blueprint (target files, exact diffs, or commands).

---

Standard Output Formats

When active, the agent presents the 100-mind deliberation in a compact, token-efficient structure:

1. Hivemind Deliberation Summary

### 👥 Collective Mind Deliberation (100 Minds / 6 Teams)

- **[Team 1 & 2 - Proposed Solution]**: [Architectural approach & core technical logic]
- **[Team 3 - Red Team Attack]**: [Critical vulnerabilities, edge cases, & flaw analysis]
- **[Team 4 - Empirical Grounding]**: [Codebase facts, log evidence, or type system verifications]
- **[Team 5 - Consensus Council]**: [Unified consensus & trade-off decisions]

2. Lean Execution Matrix (Team 6 Ops)

| Target File / Task | Responsible Team | Actions / Modifications | Risk Level |
|---|---|---|---|
| `d:/.../file.ts` | Team 2 & Team 6 | [Minimalist, precise changes] | Low |

---

Core Invariant Rules

1. Anti-Hallucination via Red Teaming: Proposals from Teams 1 & 2 must be challenged by Teams 3 & 4 before approval.
2. Pragmatic Veto Power: Team 6 holds veto power over any over-engineered or excessively bloated solution.
3. Symbolic Aggregation: Do not output lengthy individual prose for all 100 people. Aggregate insights through the 6 specialized committees for maximum density and speed.