Behavioral Core

Plaited Behavioral Core

Purpose

This skill provides comprehensive documentation for Plaited's behavioral programming paradigm—a general coordination mechanism for managing concurrent behaviors through event-driven synchronization. BP is the foundation for all Plaited applications, whether UI or non-UI.

Use this when:

• Implementing event-driven coordination with behavioral() and useBehavioral
• Designing rule composition with bThread/bSync
• Understanding event selection, blocking, and priority
• Building symbolic reasoning layers for constraint validation
• Orchestrating test runners (see useRunner in plaited/workshop/use-runner.ts)
• Implementing game logic and state machines (see plaited/src/main/tests/tic-tac-toe.spec.ts)
• Designing workflow coordinators and protocol implementations

Quick Reference

Testing: BP logic is tested with Bun tests (*.spec.ts), not browser stories. See src/main/tests/ for examples.

TypeScript LSP: Use the typescript-lsp@plaited_development-skills skill for type inference from plaited package imports.

Core Patterns

Behavioral Programs

behavioral-programs.md - Complete BP paradigm documentation

Key Capabilities:

1. Event Selection Strategy - Priority-based selection with blocking precedence
2. Rule Composition Patterns - Additive composition of independent threads
3. Predicate-Based Event Matching - Dynamic, state-dependent event handling
4. Thread Lifecycle & Runtime Management - Dynamic thread addition/removal

Core APIs:

• behavioral() - Create a behavioral program instance
• bThread() - Compose synchronization points into sequences
• bSync() - Define synchronization idioms (request, waitFor, block, interrupt)
• useBehavioral() - Factory pattern for reusable BP configurations
• useFeedback() - Register side-effect handlers
• useSnapshot() - Observe BP engine decisions

Synchronization Idioms

import { behavioral, bThread, bSync } from 'plaited'

const { trigger, bThreads, useFeedback } = behavioral()

bThreads.set({
  // Request: propose events
  requester: bThread([
    bSync({ request: { type: 'submit' } })
  ]),

  // WaitFor: pause until matching event
  listener: bThread([
    bSync({ waitFor: 'submit' }),
    bSync({ request: { type: 'confirmed' } })
  ]),

  // Block: prevent events (takes precedence)
  validator: bThread([
    bSync({ block: ({ type }) => type === 'submit' && !isValid() })
  ], true),

  // Interrupt: terminate thread on event
  cancellable: bThread([
    bSync({ request: { type: 'poll' } }),
    bSync({ waitFor: 'response' })
  ], true, { interrupt: 'cancel' })
})

Event Selection Algorithm

flowchart TD
    A[1. COLLECT candidates] --> B[Requested events become candidates]
    B --> C[2. APPLY BLOCKING]
    C --> D[Remove blocked candidates]
    D --> E[3. SELECT by PRIORITY]
    E --> F[External triggers: priority 0]
    F --> G[Thread requests: by registration order]
    G --> H{Candidates available?}
    H -->|YES| I[Execute selected event]
    H -->|NO| J[Wait for external trigger]

useBehavioral Factory Pattern

import { useBehavioral } from 'plaited'

const createProgram = useBehavioral<Events, Context>({
  publicEvents: ['START', 'STOP'],  // API whitelist

  bProgram({ trigger, bThreads, bSync, ...context }) {
    bThreads.set({
      workflow: bThread([
        bSync({ waitFor: 'START' }),
        bSync({ request: { type: 'INTERNAL' } }),
        bSync({ waitFor: 'STOP' })
      ])
    })

    return {
      START() { /* handler */ },
      INTERNAL() { /* internal only */ },
      STOP() { /* handler */ }
    }
  }
})

const publicTrigger = await createProgram({ ...context })
publicTrigger({ type: 'START' })     // Allowed
publicTrigger({ type: 'INTERNAL' })  // Throws: not in publicEvents

Non-UI Use Cases

Test Orchestration

From plaited/workshop/use-runner.ts:

const useRunner = useBehavioral<Events, Context>({
  publicEvents: ['run'],

  async bProgram({ trigger, bThreads, bThread, bSync, stories }) {
    let failed: TestResult[] = []
    let passed: TestResult[] = []

    bThreads.set({
      onCountChange: bThread([
        bSync({
          waitFor: ({ type }) => {
            if (!['test_fail', 'test_pass'].includes(type)) return false
            const remaining = stories.size - (failed.length + passed.length)
            return remaining === 1
          }
        }),
        bSync({ request: { type: 'report' } }),
        bSync({ request: { type: 'end' } })
      ], true)
    })

    return {
      test_pass({ pathname }) { passed.push(...) },
      test_fail(detail) { failed.push(...) },
      report() { /* print summary */ },
      end() { process.exit(failed.length > 0 ? 1 : 0) }
    }
  }
})

Game Logic (Tic-Tac-Toe)

See plaited/src/main/tests/tic-tac-toe.spec.ts for complete example demonstrating:

• Turn enforcement via blocking
• Square occupation tracking
• Win detection through predicate-based waiting
• Additive rule composition

Workflow Coordination

bThreads.set({
  orderWorkflow: bThread([
    bSync({ waitFor: 'orderReceived' }),
    bSync({ request: { type: 'validateOrder' } }),
    bSync({ waitFor: 'orderValid' }),
    bSync({ request: { type: 'processPayment' } }),
    bSync({ waitFor: 'paymentProcessed' }),
    bSync({ request: { type: 'shipOrder' } })
  ]),

  fraudCheck: bThread([
    bSync({ waitFor: 'processPayment' }),
    bSync({ request: { type: 'checkFraud' } }),
    bSync({ waitFor: 'fraudChecked' }),
    bSync({
      block: ({ type }) => type === 'shipOrder' && isFraudulent
    })
  ])
})

Neuro-Symbolic Reasoning

BP functions as the symbolic reasoning layer in neuro-symbolic architectures:

flowchart TB
    subgraph Neural["Neural Layer"]
        Reasoning["Agent Reasoning"]
        Actions["Proposed Actions<br/>(with confidence)"]
    end

    subgraph Symbolic["Symbolic Layer (BP)"]
        Rules["Rule Composition"]
        Constraints["Constraint Validation"]
        Selection["Event Selection"]
    end

    Actions -->|"Events with priority"| Rules
    Rules --> Constraints --> Selection
    Selection -->|"Validated Actions"| Output["System Actions"]

How BP enables symbolic reasoning:

1. Event Selection - Neural outputs become events with confidence → priority
2. Rule Composition - Add constraints without modifying neural models
3. Predicate Matching - Context-aware symbolic validation
4. Thread Lifecycle - Adaptive reasoning via runtime rule management

Actor Model Orchestration

For large-scale systems, compose multiple BP programs through an actor model:

// Orchestrator coordinates specialized workers
const createOrchestrator = useBehavioral({
  publicEvents: ['TASK_RECEIVED'],

  async bProgram({ trigger, bThreads }) {
    const analysisWorker = await createAnalysisWorker({ taskSignal, resultSignal })
    const validationWorker = await createValidationWorker({ taskSignal, resultSignal })

    resultSignal.listen('result', trigger)

    bThreads.set({
      coordinator: bThread([
        bSync({ waitFor: 'TASK_RECEIVED' }),
        bSync({ request: { type: 'ROUTE_TASK' } }),
        bSync({ waitFor: 'result' }),
        bSync({ request: { type: 'COMPLETE_TASK' } })
      ], true)
    })

    return { /* handlers */ }
  }
})

Behavioral Utilities

• useRandomEvent(e1, e2, ...) - Random event selection (template function)
• shuffleSyncs(s1, s2, ...) - Randomize sync point order (Fisher-Yates)
• isBPEvent(value) - Runtime event validation type guard
• isPlaitedTrigger(trigger) - PlaitedTrigger type guard

Related Skills

• standards - Code conventions, development standards, verification workflow
• ui-patterns - UI application of BP (bElements, templates, styling)
• typescript-lsp@plaited_development-skills - Type verification and symbol discovery