Muonroi.RuleEngine.Proliferation 1.0.0-alpha.16

Muonroi.RuleEngine.Proliferation

AI-driven neuron scenario generation and execution for the Muonroi RuleEngine — automatically proliferates test and simulation scenarios from your live rule set using Ollama, OpenAI, or Claude.

The Proliferation Engine connects to a running Muonroi.RuleEngine.Runtime instance and uses a configurable AI brain to analyse your rule set, generate NeuronScenario objects that cover untested paths, and execute them automatically via a ProliferationWorker hosted service. Scenarios, results, and rule lineage are persisted through IProliferationStore — defaulting to an in-memory store; swap to Postgres by adding Muonroi.RuleEngine.Proliferation.Persistence.

Installation

dotnet add package Muonroi.RuleEngine.Proliferation --prerelease

Quick Start

using Muonroi.Integration.Connectors.Registration;
using Muonroi.RuleEngine.Proliferation;

// IServiceTaskConnector is required by ExternalScenarioExecutor.
// AddMBuiltInConnectors() provides it. Register a standalone config
// provider so every scenario routes to the internal executor.
builder.Services.AddMBuiltInConnectors();
builder.Services.AddSingleton<IExternalProjectConfigProvider,
    StandaloneExternalProjectConfigProvider>();

// Register the AI scenario-generation stack:
// brain provider (Ollama / OpenAI / Claude per ProliferationOptions),
// prompt builder, deduplicators, budget allocator, scenario executors,
// and the ProliferationWorker hosted service.
// Defaults to InMemoryProliferationStore.
builder.Services.AddMProliferationEngine(builder.Configuration);

appsettings.json — brain selection and key tunables:

{
  "Proliferation": {
    "BrainProvider": "ollama",
    "OllamaEndpoint": "http://127.0.0.1:11434",
    "PrimaryModel": "qwen2.5-coder:14b-instruct-q5_K_M",
    "AiTimeoutSeconds": 60,
    "EnableSemanticDedup": false,
    "EnableInfraAwareBudget": false
  }
}

Once running, ProliferationWorker polls every WorkerIntervalSeconds (default 300 s), calls IRuleProliferationBrain.AnalyzeAsync, and hands new NeuronScenario objects to IScenarioExecutor.

Features

• Pluggable AI brain — built-in OllamaProliferationBrain, OpenAiProliferationBrain, and ClaudeProliferationBrain; swap via BrainProvider config key.
• Composite brain mode — chain multiple providers in parallel or sequential order via CompositeBrains + CompositeMode.
• Failure feedback loop — DefaultFailureAnalyzer drives additional scenario generation from execution failures when EnableFailureFeedback is true.
• Semantic deduplication — VectorSemanticDeduplicator (backed by OllamaEmbedder) prevents near-duplicate scenarios when EnableSemanticDedup is true.
• Hash deduplication — InputHashDeduplicator is always active as the base deduplication layer.
• Chaos mode — DefaultChaosScenarioGenerator generates adversarial edge-case scenarios when EnableChaosMode is true.
• Coverage-weighted budget — CoverageWeightedBudgetAllocator directs generation budget towards uncovered rule paths when EnableSmartBudget is true.
• Infrastructure-aware budget — InfraHealthMonitor dynamically scales the budget based on measured endpoint latency when EnableInfraAwareBudget is true.
• Continuous simulation — autonomous run loop (interval, batch size, and strategy are configurable) when EnableContinuousMode is true.
• Auto-trigger — re-proliferates automatically after rule-set changes (debounced) when EnableAutoTrigger is true.
• Multi-tenant simulation — routes scenarios per configured SimulatedTenantIds when EnableMultiTenantSimulation is true.
• External project routing — RoutingScenarioExecutor delegates to ExternalScenarioExecutor with mTLS, OAuth2, and API-key auth when a control-plane IExternalProjectConfigProvider is present.
• Natural language rule conversion — NaturalLanguageRuleConverter translates plain-text rule descriptions to structured rule JSON via the configured brain.
• Regression runner — DefaultRegressionRunner re-executes historical scenarios to guard against regressions.
• Webhook notifications — optional IWebhookNotificationService delivers lifecycle events to external consumers.
• Pluggable persistence — IProliferationStore defaults to InMemoryProliferationStore; replace with Muonroi.RuleEngine.Proliferation.Persistence for Postgres-backed durability.

Configuration

All options are bound from the "Proliferation" configuration section (ProliferationOptions.SectionName).

API Reference

Samples

• Quickstart.RuleEngine.Proliferation — end-to-end ASP.NET Core API demonstrating AddMProliferationEngine + Postgres persistence via AddMProliferationPostgres, with a stats controller reading from IProliferationStore.

Compatibility

• Target framework: net8.0
• License: Apache-2.0 (OSS)

Related Packages

• Muonroi.RuleEngine.Abstractions — contracts shared across the rule engine stack
• Muonroi.RuleEngine.Runtime — rule execution runtime that the Proliferation Engine generates scenarios against
• Muonroi.Integration.Connectors — provides IServiceTaskConnector required by ExternalScenarioExecutor

License

Apache-2.0. See LICENSE-APACHE for terms.