Skip to content

Core Harness

Component id: core

Source contract: BlockContract for core.

Overview

core provides the foundation types that every other G6 component builds on, including the generic AIBlock base class, the railway-oriented Result monad, guardrailed SearchTree, resource-bounding primitives, and structural protocols.

Launch readiness boundary

The core infrastructure test suite passes and the component is suitable as the launch foundation for typed results, pipeline execution, resource bounds, update safety, and production metrics. That does not make the full G6 product production-ready by itself. Before first-user or paid-user pilots, still verify the MCP clean install path, Docker images, staging deployment, Stripe end-to-end flow, production environment validation, and onboarding workflow against the launch checklist.

When to use:

  • Defining custom reasoning blocks by subclassing AIBlock
  • Building typed, composable pipelines with PipelineBlock and >> chaining
  • Enforcing token, time, and disk resource bounds on any pipeline execution

Good at

  • defining custom reasoning blocks by subclassing AIBlock
  • building typed, composable pipelines with PipelineBlock and >> chaining
  • enforcing token, time, and disk resource bounds on a pipeline

Not suited for

  • domain reasoning or task-specific logic — core is infrastructure, not a solver
  • standing in for end-to-end product readiness (passing core tests is not full launch readiness)

Mitigates

  • None declared.

Known Failure Modes

  • review_pending: operator_review

State Surface

  • Persistence: none
  • Queryable via: none
  • Retention: run

Extension Points

  • This harness has not opted into T3-added MCP skills.

Verification

  • Method: tier1_review_pending
  • Human review gate: False
  • Fallback block: none