codex-task-harness

Scope management for broad tasks requires freezing boundaries and defining explicit scope upfront. This task harness collapses vague briefs into bounded execution plans with defined risks, verification criteria, and deliverables to prevent uncontrolled expansion.

An execution harness structures workflow coordination by collapsing large requests into bounded scopes with risk framing and verification steps. It determines which skills to invoke, orders their execution, and defines how outputs should be structured across backend, documentation, and research domains.

Skill routing for cross-cutting requests is handled by evaluating the broad request scope and generating an orchestration plan. The harness determines which skills to invoke, sequences the minimal execution path, and ensures front-end contracts and backend integrations are preserved.

Research analysis and academic writing planning are supported use cases for this harness. It transforms broad paper or research requests into bounded deliverables with explicit verification criteria, risk assessment, and structured artifact generation like specifications and checklists.

Defining verification criteria requires risk assessment and failure mode identification during scope freezing. The harness generates explicit validation steps and recommended tests alongside deliverables to ensure clear verification paths and minimal execution paths are established.

Bounded execution harnesses are not ideal for simple, single-scope tasks that fit within one specific skill's domain. They are designed for large, cross-cutting requests that risk drifting beyond a single skill's scope and require complex workflow coordination.