Plan-Execute Separation
An agent orchestration paradigm in which the model first generates a complete execution plan, which an engineering framework then schedules and executes step by step. Plans can be reviewed, cached and reused, making the approach suitable for high-risk tasks that require human confirmation.
Timeline (last 90 days)
The plan-execute separation paradigm first has the model produce a complete list of steps in one go, and then the harness schedules execution step by step; the advantages are that the plan can be reviewed by humans, cached, and reused, while the disadvantage is that it needs to be re-planned when unexpected situations arise midway.
Game development involves many high-risk changes, so a hybrid of ReAct and plan-execute separation is usually adopted: low-risk tasks go through ReAct to improve efficiency, while when core systems or resource deletion are involved, the plan is first frozen and execution waits for human confirmation.
All Facts (2)
The plan-execute separation paradigm first has the model produce a complete list of steps in one go, and then the harness schedules execution step by step; the advantages are that the plan can be reviewed by humans, cached, and reused, while the disadvantage is that it needs to be re-planned when unexpected situations arise midway.
50%UnverifiedGame development involves many high-risk changes, so a hybrid of ReAct and plan-execute separation is usually adopted: low-risk tasks go through ReAct to improve efficiency, while when core systems or resource deletion are involved, the plan is first frozen and execution waits for human confirmation.
50%