Agent Ecosystem

Perex

The Prismatic Platform operates through a distributed network of 552 autonomous agents, each a specialized Claude-powered entity with precisely defined responsibilities, tool access, and escalation protocols. This agent ecosystem represents the platform's execution layer: where strategic intent transforms into operational reality through coordinated AI action.

Every agent conforms to the AIAD (AI Agent Definition) standard, ensuring predictable behavior, traceable decision-making, and governed coordination across 16 operational domains through a five-level authority hierarchy.

Key Agent Resources

552+ Autonomous Agents • 16 Operational Domains • 5 Authority Levels • AIAD Standard Compliance

The platform's agent ecosystem enables specialized deep expertise while maintaining broad cross-domain coordination through structured protocols, shared epistemic context, and NO MERCY, NO DOUBTS doctrine enforcement across all autonomous operations.

Agent Authority Hierarchy

Platform Statistics

• Total Agents: 552 autonomous entities with Claude-powered reasoning
• Operational Domains: 16 specialized areas of expertise
• Authority Levels: 5-tier hierarchy from apex coordinators to utility functions
• Standard Compliance: AIAD specification across all agents
• Communication: Structured protocols with epistemic framework integration
• Governance: NO MERCY, NO DOUBTS doctrine enforcement
• Platform Version: Prismatic v8.0 (Gen 19 Ecosystem Expansion)

Quick Navigation

• AIAD Standard - AI Agent Definition specification and compliance framework
• Agent Commands - 217 operational commands powered by agent coordination
• Platform Architecture - System design enabling multi-agent coordination
• Applications Catalog - 625 applications leveraging agent intelligence
• Developer Portal - Tools for building and integrating with agents

Explore Platform

Want to Know More?

Dive deeper into the AIAD standard that governs agent behavior, explore the operational domains below, or review the authority framework that enables coordinated autonomous action.

About Agents: Authority Hierarchy • AIAD Standard • Domain Organization (see sections below)

Abstract

This documentation provides a comprehensive reference to the Prismatic Platform's agent ecosystem, a multi-agent system comprising 552 autonomous entities organized across 16 operational domains. Each agent operates within the AIAD (AI Agent Definition) standard, ensuring consistent specification, predictable behavior, and traceable decision-making. The ecosystem implements hierarchical authority through five operational levels (L1-L5), enabling both specialized deep expertise and broad cross-domain coordination. Agents communicate through structured protocols, share context through the platform's epistemic framework (NABLA Infinity), and execute under the NO MERCY, NO DOUBTS doctrine that governs all platform operations. This index serves as the primary navigation point for exploring individual agent specifications, understanding domain organization, and comprehending the coordination mechanisms that enable 552 autonomous entities to function as a coherent intelligence platform.

1. Introduction

1.1 The Multi-Agent Paradigm

Traditional software systems execute through deterministic code paths: input flows through functions, state changes occur, output emerges. The Prismatic Platform operates differently. Rather than encoding intelligence in static code, the platform distributes cognition across hundreds of specialized agents, each capable of reasoning about its domain, coordinating with peers, and adapting to novel situations.

This architectural choice reflects a fundamental insight: enterprise intelligence tasks are too varied, too context-dependent, and too rapidly evolving for monolithic implementations. An OSINT investigation requires different reasoning than a security audit, which differs from a compliance assessment, which differs from code generation. By instantiating specialized agents for each domain, the platform achieves depth of expertise impossible in generalist systems.

1.2 The AIAD Standard

Every agent in the ecosystem conforms to the AIAD (AI Agent Definition) standard, a specification format that ensures:

• Predictable Structure: All agents define identity, purpose, authority level, tool access, and operational constraints in a consistent YAML schema
• Traceable Decisions: Agent reasoning follows documented protocols, enabling audit and verification
• Governed Behavior: Policies embedded in agent specifications enforce compliance with platform doctrine
• Composable Coordination: Standardized interfaces enable agents to delegate, escalate, and collaborate

The AIAD standard transforms agents from opaque AI black boxes into transparent, governable system components.

1.3 Scope of This Documentation

This section documents all 552 agents currently deployed in the Prismatic Platform. Each agent page provides:

• Executive summary and operational mandate
• Technical architecture and implementation details
• Tool access and capability matrix
• Authority framework and escalation paths
• Integration patterns with related agents
• Usage examples and invocation patterns

2. Ecosystem Architecture

2.1 Hierarchical Authority Model

Agents operate within a five-level authority hierarchy that balances specialization with coordination:

Higher-level agents can direct lower-level agents within their domain. Cross-domain coordination flows through L1-L2 agents or explicit delegation protocols.

2.2 Operational Domains

The 552 agents distribute across 16 operational domains, each addressing a distinct capability area:

2.3 Communication Patterns

Agents communicate through four primary patterns:

Direct Delegation: Higher-authority agents issue directives to lower-authority agents within their domain. Directives carry mandatory execution semantics.

Peer Consultation: Same-level agents exchange information and request expertise. Consultation is advisory, not binding.

Escalation: Agents escalate unresolvable situations to higher-authority agents or human operators. Escalation includes full context and recommended actions.

Broadcast: Platform-wide events (quality alerts, security incidents, evolution triggers) propagate to subscribed agents through PubSub channels.

3. Agent Categories

3.1 Supreme Agents

The three L1 apex agents hold platform-wide authority:

• ARCHER SUPREME: Impossible mission execution, crisis intervention, cross-domain orchestration
• Supreme Coordinator: Strategic planning, resource allocation, priority arbitration
• Supreme Commander: Emergency response, platform-level decisions, doctrine enforcement

These agents can override standard authority boundaries when platform integrity or mission success requires unified command.

3.2 Domain Commanders

Each operational domain has designated L2 commanders responsible for:

• Setting domain strategy and priorities
• Coordinating agents within the domain
• Interfacing with other domain commanders
• Escalating cross-domain conflicts to L1

Examples: Blue Commander, Quality Commander, CER Compliance Commander.

3.3 Specialist Agents

L3-L4 specialists provide deep expertise in narrow domains:

• Analysis Specialists: Bayesian Analyst, Corporate Governance Analyst
• Technical Specialists: Elixir Architect, Database Specialist
• Investigation Specialists: Czech Registry Investigator, Email Intelligence Specialist

3.4 Tactical Operatives

L4-L5 operatives execute atomic tasks:

• Execution Agents: Focused task agents for rapid, atomic operations
• Collection Agents: Crawler Development Specialist, Data Sync Specialist
• Monitoring Agents: Deployment Health Monitor, Ecosystem Biologist Coordinator

4. Epistemic Framework Integration

4.1 NABLA Infinity Compliance

All agents operate within the NABLA Infinity epistemic framework, adhering to seven non-negotiable axioms:

1. Signal Plurality: No agent acts on single-source information for critical decisions
2. Contradiction Evidence: Conflicting signals are preserved, not discarded
3. Absence Informative: Missing data is tracked as meaningful signal
4. Time Decay: All beliefs carry timestamps and decay functions
5. Unknown Valid: "I don't know" is a legitimate and valued response
6. Source Independence: Weight increases with source diversity
7. Provenance Mandatory: All beliefs trace to originating evidence

4.2 Trinity Gate Validation

Before executing high-stakes decisions, agents must pass the Trinity Gate:

1. Structural Consistency: Graph-theoretic validation of belief network coherence
2. Logical Consistency: Rule-based verification of inference chains
3. Formal Necessity: Modal logic validation with optional Lean4 proofs

Confidence must reach domain-appropriate thresholds (0.95 for critical, 0.80 for standard, 0.60 for exploratory) before execution proceeds.

5. Color Team Security Agents

Twenty specialized agents operate across six color teams for adversarial-defensive security:

All color team operations execute in sandboxed environments with synthetic data only.

6. Evolution and Self-Improvement

6.1 SEADF Integration

The Self-Evolving Autonomous Development Framework enables agents to improve over time:

• Pattern Learning: Successful execution patterns are extracted and generalized
• Capability Expansion: Agents can request new tools based on operational needs
• Performance Optimization: Execution metrics drive continuous tuning
• Cross-Pollination: Effective patterns propagate across agents

6.2 Generation Tracking

The platform tracks evolutionary progress through generations. Current state: Generation 18 with 0.999 apex fitness. Each generation represents accumulated improvements across the agent ecosystem.

7. Governance and Compliance

7.1 NO MERCY, NO DOUBTS Doctrine

All agents operate under the universal platform doctrine:

NO MERCY: Zero tolerance for incomplete implementations, untested code, quality violations, or deferred problems.

NO DOUBTS: Full investigation before action, decisive execution once committed, evidence-backed claims only, verified results without assumptions.

7.2 Audit and Traceability

Every agent action generates audit records including:

• Agent identity and authority level
• Input context and decision factors
• Tool invocations and their results
• Output produced and confidence levels
• Execution duration and resource consumption

8. Navigation Guide

8.1 By Domain

Select a domain to explore its agents:

• Supreme Agents - Apex coordinators (3 agents)
• Strategic Agents - Cross-domain planning (47 agents)
• OSINT Agents - Intelligence gathering (62 agents)
• Security Agents - Attack surface management (38 agents)
• Quality Agents - Code and process quality (45 agents)
• Development Agents - Code generation and maintenance (58 agents)

8.2 By Authority Level

• L1 Apex Agents - 3 agents with platform-wide authority
• L2 Strategic Commanders - 28 domain-level coordinators
• L3 Operational Specialists - 87 complex task executors
• L4 Tactical Operatives - 156 focused task agents
• L5 Support Agents - 156 utility and monitoring agents

8.3 Alphabetical Index

All 552 agents are listed alphabetically in the main agent listing below.

9. Related Documentation

• Commands - CLI commands that invoke agents
• Applications - Umbrella apps hosting agent runtime
• Architecture - Technical infrastructure
• AIAD Standard - Agent specification format
• NABLA Infinity - Epistemic framework

10. Conclusion

The Prismatic Platform's agent ecosystem represents a new paradigm in enterprise software: distributed cognition through specialized autonomous agents, governed by rigorous epistemic standards, executing under uncompromising quality doctrine. Each of the 552 agents documented in this section contributes unique capabilities to the collective intelligence of the platform.

Browse the agent listings below to explore individual specifications, or use the domain and level filters to navigate by organizational structure. Every agent page provides the depth of documentation necessary to understand, invoke, and integrate with the platform's autonomous capabilities.

552 autonomous agents. 16 operational domains. One unified intelligence platform.

Connect & Contribute

Created by Tomas Korcak (korczis) | Open Source under GHL

• GitHub | GitLab | LinkedIn | Contact
• Developer Portal | Architecture | Meet the Creator