Skip to main content
Jira progress: loading…

WTE

Workflow Test Engine

1. Purpose​

The Workflow Test Engine is the assurance engine that validates Workflow Definitions and isolated Workflow Executions against governed functional, temporal, authority, gate, security, side-effect, and provenance requirements. It supports static conformance checks, scenario execution, simulation, dry-run, replay, shadow testing, regression testing, and migration validation. It must preserve test evidence and must not permit authoritative side effects from non-authoritative test modes.

  • Workflow definitions must be validated for completeness, unreachable states, dead ends, unbounded loops, missing compensation/escalation, gate and authority coverage, task contracts, event correlation, temporal correctness, tenant isolation, and provenance (/cdf/constitutional-runtime-framework-8, §23.79.12).
  • Workflow execution supports explicit test, simulation, replay, shadow, and dry-run modes; simulation/dry-run cannot create authoritative side effects (/cdf/constitutional-runtime-framework-8, §23.78.6).
  • Replays must use the historical definition, context, events, human decisions, agent/tool outputs, policies, runtime bundle, and temporal state, while remaining isolated and side-effect controlled (/cdf/constitutional-runtime-framework-8, §23.80.16).
  • Workflow transitions and task completion have explicit gate, authority, evidence, temporal, side-effect, and provenance requirements (/cdf/constitutional-runtime-framework-8, §23.80–23.81).
  • Workflow Runtime / Engine: executes and coordinates governed workflow instances.
  • TrustGate Scheduler / Orchestrator: decides when and how background/asynchronous workflows and jobs run; it explicitly does not make assurance decisions (/trustgate/trustgate-engines/trustgate-scheduler-orchestrator-engine).
  • Workflow Test Engine: evaluates whether definitions and executions conform to expected workflow semantics and assurance rules, ideally in isolated test/simulation/replay environments.

A useful specification could include:

  1. Purpose and non-goals

    • Conformance testing and assurance of workflows.
    • Not a production scheduler, authority resolver, or primary workflow runtime.
  2. Inputs

    • Workflow Definition and version
    • Workflow Profile
    • Runtime Context / context fingerprint
    • Test scenario / fixture
    • Event sequences, timer state, task/participant inputs
    • Expected transitions, gate outcomes, outputs, evidence, and side-effect assertions
    • Runtime Bundle and policy versions
  3. Test modes

    • Static definition validation
    • Unit testing of tasks, predicates, and transitions
    • End-to-end execution tests
    • Simulation and dry-run
    • Historical replay / deterministic regression testing
    • Shadow execution
    • Migration and compatibility testing across workflow-definition versions
  4. Required checks

    • Reachability, dead ends, loops, forks/joins
    • Gate and authority coverage
    • Task input/output contract validity
    • Timer and event-wait behavior
    • Compensation, escalation, suspension, and resumption paths
    • Tenant/security isolation
    • Side-effect suppression or reconciliation
    • Provenance/evidence completeness
    • Determinism under equivalent inputs and histories
  5. Outputs

    • Pass/fail/indeterminate result
    • Failed invariant or contract
    • Execution trace and transition history
    • Expected-versus-actual differences
    • Definition/version/context/runtime-bundle fingerprints
    • Test evidence suitable for audit and release assurance
  6. Integration points

    • HECATE for definition and runtime validation
    • Workflow Definition/Profile resolution
    • Workflow Runtime for isolated execution
    • Scheduler only for scheduling periodic/regression suites, if needed
    • CI/CD or constitutional compilation outputs, since workflows may compile into conformance fixtures (/cdf/constitutional-runtime-framework-8, §23.79.13)



GitHub RepoRequest for Change (RFC)