Prismatic

Overview

Prismatic is the core component of the Prismatic Platform's umbrella architecture. As the main application with 1,640 source files, it serves as the central coordination hub for OSINT operations, investigation workflows, agent orchestration, and cross-domain intelligence fusion. Every other application in the 99-app umbrella either depends on Prismatic for foundational services or coordinates through its public API.

The application embodies the platform's core philosophy: OTP-first design with stateful processes, supervision trees for fault tolerance, and pure functions at the computational boundaries. It houses the primary Mix tasks that drive platform operations, the facade modules that external applications consume, and the coordination logic that ties the entire ecosystem together.

As the largest application in the umbrella, Prismatic bears a unique architectural responsibility. It must provide stable, well-documented public APIs that dozens of downstream applications depend on, while continuously evolving to support new intelligence capabilities and operational workflows. This tension between stability and evolution is managed through a strict facade module pattern where public APIs are versioned and backward-compatible, while internal implementation details can change freely.

Architecture

Core Responsibilities

OSINT Coordination

Prismatic orchestrates open-source intelligence operations across 250+ providers with priority scheduling, deduplication, and result aggregation:

The OSINT coordination layer manages the complexity of querying hundreds of intelligence sources with varying rate limits, response formats, and reliability characteristics. The Priority Scheduler uses a weighted scoring system that considers query urgency (investigation priority), source quality (historical accuracy and relevance), and cost (API credits or rate limit budget) to determine query ordering. This ensures that high-priority investigations receive intelligence data first, while routine monitoring operates within sustainable rate limit budgets.

# OSINT coordination through Prismatic facade
{:ok, results} = Prismatic.investigate("example.com", [
  providers: [:shodan, :censys, :certificate_transparency],
  depth: 3,
  parallel: true,
  timeout: 30_000
])

Investigation Workflows

End-to-end investigation lifecycle management from initial lead through evidence collection to final reporting:

Lead --> Triage --> Collection --> Analysis --> Correlation --> Report --> Archive
  |         |           |            |              |             |          |
Target   Priority   Multi-source  Pattern      Cross-domain  Structured  Long-term
Input    + Scope    OSINT Queries  Matching     Entity Links  Output      Storage

The investigation workflow engine supports both fully automated and analyst-guided investigation modes. In automated mode, the system follows pre-defined investigation templates that specify which sources to query, what analysis to perform, and how to correlate results. In guided mode, analysts make decisions at each phase while the system handles the operational complexity of source queries, data normalization, and result aggregation.

Agent Orchestration

Multi-agent task coordination with dependency resolution, parallel execution, and result aggregation:

Cross-Domain Fusion

Intelligence fusion engine combining data from network, financial, legal, and social domains into unified intelligence products:

The cross-domain fusion engine is the analytical heart of the platform. It takes entity data from five independent domains and identifies connections, patterns, and anomalies that are invisible within any single domain. A domain's network infrastructure (discovered through DNS and certificate scans) is correlated with its corporate ownership (from business registries) and financial relationships (from AML databases), producing a unified entity profile that captures the full complexity of the target.

Facade Modules

Prismatic exposes its functionality through well-defined facade modules that serve as the public API:

The Prismatic API auto-discovers these facade modules at boot time using Elixir introspection (Code.fetch_docs/1, Module.__info__/1) and exposes them as REST endpoints. This means that every function added to a facade module automatically becomes available through the REST API without additional configuration.

Mix Tasks

Prismatic houses the majority of platform Mix tasks that drive development and operations:

Testing

mix test apps/prismatic/test
mix test apps/prismatic/test --cover

Integration Points

NABLA Compliance

As the central coordination hub, Prismatic enforces NABLA compliance across all intelligence operations. The Result Aggregator implements Signal Plurality by combining intelligence from multiple independent sources with explicit confidence weights. The Deduplication Engine satisfies the Contradiction Preservation axiom by logging contradictory results from different providers rather than silently discarding one. Investigation workflows carry full provenance through every phase, recording the sources queried, the analysis performed, and the decisions made, satisfying the Provenance Mandatory axiom. Cross-domain fusion operations require Trinity Gate passage for high-confidence assertions that span multiple intelligence domains.

Related Components

• Prismatic Web - LiveView dashboards and user interface
• Prismatic API - REST API gateway with auto-discovery
• Prismatic Storage - Data persistence layer
• Prismatic Claude - Claude AI session management
• Prismatic Safety - Quality floor guardian and autoheal

Related Agents

• Elixir Architect -- Ensures the core application follows OTP-first design with proper supervision topology
• Architecture Review Specialist -- Reviews cross-domain coordination architecture and facade module design
• Consolidation Architect -- Data deduplication and entity resolution across OSINT provider results
• GitLab Strategic Coordinator -- Strategic coordination of development efforts across the core application

Related Capabilities

• Intelligence Synthesis -- Cross-domain intelligence fusion engine combining network, financial, legal, and social data
• NABLA Axioms -- Confidence-scored result aggregation compliant with epistemic axioms
• Autonomous Self-Healing -- Autoheal and autoevolve cycles maintaining platform health

Connect & Contribute

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

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